Euanyu/combined_bionlp_task_dataset_model_cards

id stringlengths 6 113	author stringlengths 2 36	task_category stringclasses 42 values	tags listlengths 1 4.05k	created_time timestamp[ns, tz=UTC]date 2022-03-02 23:29:04 2025-04-10 08:38:38	last_modified stringdate 2020-05-14 13:13:12 2025-04-19 04:15:39	downloads int64 0 118M	likes int64 0 4.86k	README stringlengths 30 1.01M	matched_bigbio_names listlengths 1 8 ⌀	is_bionlp stringclasses 3 values	model_cards stringlengths 0 1M	metadata stringlengths 2 698k	source stringclasses 2 values	matched_task listlengths 1 10 ⌀	__index_level_0__ int64 0 46.9k
IAmSkyDra/BART_Translation_Finetune_v0	IAmSkyDra	text2text-generation	[ "transformers", "safetensors", "mbart", "text2text-generation", "generated_from_trainer", "base_model:vinai/bartpho-syllable", "base_model:finetune:vinai/bartpho-syllable", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2025-01-29T07:54:22Z	2025-01-29T09:14:59+00:00	17	0	--- base_model: vinai/bartpho-syllable library_name: transformers license: mit metrics: - sacrebleu tags: - generated_from_trainer model-index: - name: BART_Translation_Finetune_v0 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # BART_Translation_Finetune_v0 This model is a fine-tuned version of [vinai/bartpho-syllable](https://huggingface.co/vinai/bartpho-syllable) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.4390 - Sacrebleu: 6.4288 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 100 - eval_batch_size: 100 - seed: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 15 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Sacrebleu \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:---------:\| \| 0.7274 \| 1.0 \| 468 \| 0.6396 \| 1.4512 \| \| 0.628 \| 2.0 \| 936 \| 0.5901 \| 2.1680 \| \| 0.5744 \| 3.0 \| 1404 \| 0.5431 \| 3.0785 \| \| 0.5348 \| 4.0 \| 1872 \| 0.5141 \| 3.9172 \| \| 0.5044 \| 5.0 \| 2340 \| 0.4905 \| 4.3428 \| \| 0.4773 \| 6.0 \| 2808 \| 0.4758 \| 4.8562 \| \| 0.4575 \| 7.0 \| 3276 \| 0.4647 \| 5.2799 \| \| 0.4372 \| 8.0 \| 3744 \| 0.4572 \| 5.5700 \| \| 0.429 \| 9.0 \| 4212 \| 0.4499 \| 5.8456 \| \| 0.4118 \| 10.0 \| 4680 \| 0.4455 \| 6.0842 \| \| 0.4056 \| 11.0 \| 5148 \| 0.4424 \| 6.2403 \| \| 0.3924 \| 12.0 \| 5616 \| 0.4403 \| 6.2796 \| \| 0.3858 \| 13.0 \| 6084 \| 0.4396 \| 6.3191 \| \| 0.386 \| 14.0 \| 6552 \| 0.4389 \| 6.4120 \| \| 0.3809 \| 15.0 \| 7020 \| 0.4390 \| 6.4288 \| ### Framework versions - Transformers 4.48.1 - Pytorch 2.5.1+cu124 - Datasets 3.2.0 - Tokenizers 0.21.0	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # BART_Translation_Finetune_v0 This model is a fine-tuned version of [vinai/bartpho-syllable](https://huggingface.co/vinai/bartpho-syllable) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.4390 - Sacrebleu: 6.4288 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 100 - eval_batch_size: 100 - seed: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 15 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Sacrebleu \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:---------:\| \| 0.7274 \| 1.0 \| 468 \| 0.6396 \| 1.4512 \| \| 0.628 \| 2.0 \| 936 \| 0.5901 \| 2.1680 \| \| 0.5744 \| 3.0 \| 1404 \| 0.5431 \| 3.0785 \| \| 0.5348 \| 4.0 \| 1872 \| 0.5141 \| 3.9172 \| \| 0.5044 \| 5.0 \| 2340 \| 0.4905 \| 4.3428 \| \| 0.4773 \| 6.0 \| 2808 \| 0.4758 \| 4.8562 \| \| 0.4575 \| 7.0 \| 3276 \| 0.4647 \| 5.2799 \| \| 0.4372 \| 8.0 \| 3744 \| 0.4572 \| 5.5700 \| \| 0.429 \| 9.0 \| 4212 \| 0.4499 \| 5.8456 \| \| 0.4118 \| 10.0 \| 4680 \| 0.4455 \| 6.0842 \| \| 0.4056 \| 11.0 \| 5148 \| 0.4424 \| 6.2403 \| \| 0.3924 \| 12.0 \| 5616 \| 0.4403 \| 6.2796 \| \| 0.3858 \| 13.0 \| 6084 \| 0.4396 \| 6.3191 \| \| 0.386 \| 14.0 \| 6552 \| 0.4389 \| 6.4120 \| \| 0.3809 \| 15.0 \| 7020 \| 0.4390 \| 6.4288 \| ### Framework versions - Transformers 4.48.1 - Pytorch 2.5.1+cu124 - Datasets 3.2.0 - Tokenizers 0.21.0	{"base_model": "vinai/bartpho-syllable", "library_name": "transformers", "license": "mit", "metrics": ["sacrebleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "BART_Translation_Finetune_v0", "results": []}]}	task	[ "TRANSLATION" ]	44,529
zedfum/arman-longformer-8k-finetuned-ensani	zedfum	summarization	[ "transformers", "pytorch", "safetensors", "pegasus", "text2text-generation", "summarization", "fa", "dataset:zedfum/long-summarization-persian", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-06-19T07:32:56Z	2023-07-18T16:31:10+00:00	68	2	--- datasets: - zedfum/long-summarization-persian language: - fa pipeline_tag: summarization --- <!-- Provide a quick summary of what the model is/does. --> Persian Long Text Summarization This Model Accept 8K Tokens ## ⚡️ Quickstart ``` from transformers import AutoTokenizer from transformers import pipeline summarizer = pipeline("summarization", model="zedfum/arman-longformer-8k-finetuned-ensani", tokenizer="zedfum/arman-longformer-8k-finetuned-ensani" , device=0) text_to_summarize="" summarizer(text_to_summarize, min_length=5, max_length=512, truncation=True) ```	null	Non_BioNLP	<!-- Provide a quick summary of what the model is/does. --> Persian Long Text Summarization This Model Accept 8K Tokens ## ⚡️ Quickstart ``` from transformers import AutoTokenizer from transformers import pipeline summarizer = pipeline("summarization", model="zedfum/arman-longformer-8k-finetuned-ensani", tokenizer="zedfum/arman-longformer-8k-finetuned-ensani" , device=0) text_to_summarize="" summarizer(text_to_summarize, min_length=5, max_length=512, truncation=True) ```	{"datasets": ["zedfum/long-summarization-persian"], "language": ["fa"], "pipeline_tag": "summarization"}	task	[ "SUMMARIZATION" ]	44,530
datlaaaaaaa/81d3479a-ff61-4e88-8c7b-3c3d80ee47bf	datlaaaaaaa	null	[ "peft", "safetensors", "llama", "axolotl", "generated_from_trainer", "base_model:unsloth/llama-3-8b", "base_model:adapter:unsloth/llama-3-8b", "license:llama3", "8-bit", "bitsandbytes", "region:us" ]	2025-01-18T12:15:07Z	2025-01-18T12:31:41+00:00	1	0	--- base_model: unsloth/llama-3-8b library_name: peft license: llama3 tags: - axolotl - generated_from_trainer model-index: - name: 81d3479a-ff61-4e88-8c7b-3c3d80ee47bf results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: unsloth/llama-3-8b bf16: auto chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 1b99c07a98e6036a_train_data.json ds_type: json format: custom path: /workspace/input_data/1b99c07a98e6036a_train_data.json type: field_input: choices field_instruction: input_translation field_output: input format: '{instruction} {input}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null deepspeed: null early_stopping_patience: null eval_max_new_tokens: 128 eval_table_size: null evals_per_epoch: 1 flash_attention: true fp16: null fsdp: null fsdp_config: null gradient_accumulation_steps: 4 gradient_checkpointing: true gradient_clipping: 1.0 group_by_length: false hub_model_id: datlaaaaaaa/81d3479a-ff61-4e88-8c7b-3c3d80ee47bf hub_repo: null hub_strategy: end hub_token: null learning_rate: 5.0e-05 load_in_4bit: true load_in_8bit: true local_rank: null logging_steps: 1 lora_alpha: 16 lora_dropout: 0.05 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 8 lora_target_linear: true lr_scheduler: cosine max_steps: 200 micro_batch_size: 2 mlflow_experiment_name: /tmp/1b99c07a98e6036a_train_data.json model_type: AutoModelForCausalLM num_epochs: 1 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false saves_per_epoch: 1 sequence_len: 1024 strict: false tf32: false tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.05 wandb_entity: null wandb_mode: online wandb_name: 582a5157-bb3b-4681-835d-d17ab3d65f4d wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 582a5157-bb3b-4681-835d-d17ab3d65f4d warmup_steps: 5 weight_decay: 0.01 xformers_attention: true ``` </details><br> # 81d3479a-ff61-4e88-8c7b-3c3d80ee47bf This model is a fine-tuned version of [unsloth/llama-3-8b](https://huggingface.co/unsloth/llama-3-8b) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.2841 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 8 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 5 - training_steps: 200 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:------:\|:----:\|:---------------:\| \| 0.2444 \| 0.7048 \| 200 \| 0.2841 \| ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: unsloth/llama-3-8b bf16: auto chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 1b99c07a98e6036a_train_data.json ds_type: json format: custom path: /workspace/input_data/1b99c07a98e6036a_train_data.json type: field_input: choices field_instruction: input_translation field_output: input format: '{instruction} {input}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null deepspeed: null early_stopping_patience: null eval_max_new_tokens: 128 eval_table_size: null evals_per_epoch: 1 flash_attention: true fp16: null fsdp: null fsdp_config: null gradient_accumulation_steps: 4 gradient_checkpointing: true gradient_clipping: 1.0 group_by_length: false hub_model_id: datlaaaaaaa/81d3479a-ff61-4e88-8c7b-3c3d80ee47bf hub_repo: null hub_strategy: end hub_token: null learning_rate: 5.0e-05 load_in_4bit: true load_in_8bit: true local_rank: null logging_steps: 1 lora_alpha: 16 lora_dropout: 0.05 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 8 lora_target_linear: true lr_scheduler: cosine max_steps: 200 micro_batch_size: 2 mlflow_experiment_name: /tmp/1b99c07a98e6036a_train_data.json model_type: AutoModelForCausalLM num_epochs: 1 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false saves_per_epoch: 1 sequence_len: 1024 strict: false tf32: false tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.05 wandb_entity: null wandb_mode: online wandb_name: 582a5157-bb3b-4681-835d-d17ab3d65f4d wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 582a5157-bb3b-4681-835d-d17ab3d65f4d warmup_steps: 5 weight_decay: 0.01 xformers_attention: true ``` </details><br> # 81d3479a-ff61-4e88-8c7b-3c3d80ee47bf This model is a fine-tuned version of [unsloth/llama-3-8b](https://huggingface.co/unsloth/llama-3-8b) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.2841 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 8 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 5 - training_steps: 200 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:------:\|:----:\|:---------------:\| \| 0.2444 \| 0.7048 \| 200 \| 0.2841 \| ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1	{"base_model": "unsloth/llama-3-8b", "library_name": "peft", "license": "llama3", "tags": ["axolotl", "generated_from_trainer"], "model-index": [{"name": "81d3479a-ff61-4e88-8c7b-3c3d80ee47bf", "results": []}]}	task	[ "TRANSLATION" ]	44,531
navjordj/flan-t5-large_en-no	navjordj	text2text-generation	[ "transformers", "pytorch", "t5", "text2text-generation", "generated_from_trainer", "en", "no", "dataset:bible_para", "license:apache-2.0", "model-index", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2022-12-02T19:55:21Z	2022-12-03T15:33:10+00:00	16	0	--- datasets: - bible_para language: - en - 'no' license: apache-2.0 metrics: - bleu tags: - generated_from_trainer model-index: - name: flan-t5-large_en-no results: - task: type: translation name: Translation dataset: name: bible_para en-no type: bible_para config: en-no split: train args: en-no metrics: - type: bleu value: 36.7184 name: Bleu --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # flan-t5-large_en-no This model is a fine-tuned version of [google/flan-t5-large](https://huggingface.co/google/flan-t5-large) on the bible_para en-no dataset. It achieves the following results on the evaluation set: - Loss: 0.6335 - Bleu: 36.7184 - Gen Len: 64.6249 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5.0 ### Training results ### Framework versions - Transformers 4.26.0.dev0 - Pytorch 1.13.0 - Datasets 2.7.1 - Tokenizers 0.13.2	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # flan-t5-large_en-no This model is a fine-tuned version of [google/flan-t5-large](https://huggingface.co/google/flan-t5-large) on the bible_para en-no dataset. It achieves the following results on the evaluation set: - Loss: 0.6335 - Bleu: 36.7184 - Gen Len: 64.6249 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5.0 ### Training results ### Framework versions - Transformers 4.26.0.dev0 - Pytorch 1.13.0 - Datasets 2.7.1 - Tokenizers 0.13.2	{"datasets": ["bible_para"], "language": ["en", "no"], "license": "apache-2.0", "metrics": ["bleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "flan-t5-large_en-no", "results": [{"task": {"type": "translation", "name": "Translation"}, "dataset": {"name": "bible_para en-no", "type": "bible_para", "config": "en-no", "split": "train", "args": "en-no"}, "metrics": [{"type": "bleu", "value": 36.7184, "name": "Bleu"}]}]}]}	task	[ "TRANSLATION" ]	44,532
dabraldeepti25/legal-ft-v0-updated	dabraldeepti25	sentence-similarity	[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:156", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss", "arxiv:1908.10084", "arxiv:2205.13147", "arxiv:1705.00652", "base_model:Snowflake/snowflake-arctic-embed-l", "base_model:finetune:Snowflake/snowflake-arctic-embed-l", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	2025-02-18T22:13:53Z	2025-02-18T22:14:27+00:00	13	0	--- base_model: Snowflake/snowflake-arctic-embed-l library_name: sentence-transformers metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:156 - loss:MatryoshkaLoss - loss:MultipleNegativesRankingLoss widget: - source_sentence: What significant change occurred in the AI landscape regarding models surpassing GPT-4 in the past twelve months? sentences: - 'Except... you can run generated code to see if it’s correct. And with patterns like ChatGPT Code Interpreter the LLM can execute the code itself, process the error message, then rewrite it and keep trying until it works! So hallucination is a much lesser problem for code generation than for anything else. If only we had the equivalent of Code Interpreter for fact-checking natural language! How should we feel about this as software engineers? On the one hand, this feels like a threat: who needs a programmer if ChatGPT can write code for you?' - 'The GPT-4 barrier was comprehensively broken In my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t? I’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.' - 'If you think about what they do, this isn’t such a big surprise. The grammar rules of programming languages like Python and JavaScript are massively less complicated than the grammar of Chinese, Spanish or English. It’s still astonishing to me how effective they are though. One of the great weaknesses of LLMs is their tendency to hallucinate—to imagine things that don’t correspond to reality. You would expect this to be a particularly bad problem for code—if an LLM hallucinates a method that doesn’t exist, the code should be useless.' - source_sentence: How does Claude enable users to interact with applications created through its interface? sentences: - 'OpenAI made GPT-4o free for all users in May, and Claude 3.5 Sonnet was freely available from its launch in June. This was a momentus change, because for the previous year free users had mostly been restricted to GPT-3.5 level models, meaning new users got a very inaccurate mental model of what a capable LLM could actually do. That era appears to have ended, likely permanently, with OpenAI’s launch of ChatGPT Pro. This $200/month subscription service is the only way to access their most capable model, o1 Pro. Since the trick behind the o1 series (and the future models it will undoubtedly inspire) is to expend more compute time to get better results, I don’t think those days of free access to the best available models are likely to return.' - 'We already knew LLMs were spookily good at writing code. If you prompt them right, it turns out they can build you a full interactive application using HTML, CSS and JavaScript (and tools like React if you wire up some extra supporting build mechanisms)—often in a single prompt. Anthropic kicked this idea into high gear when they released Claude Artifacts, a groundbreaking new feature that was initially slightly lost in the noise due to being described half way through their announcement of the incredible Claude 3.5 Sonnet. With Artifacts, Claude can write you an on-demand interactive application and then let you use it directly inside the Claude interface. Here’s my Extract URLs app, entirely generated by Claude:' - "Industry’s Tardy Response to the AI Prompt Injection Vulnerability on RedMonk\ \ Conversations\n\n\nPosted 31st December 2023 at 11:59 pm · Follow me on Mastodon\ \ or Twitter or subscribe to my newsletter\n\n\nMore recent articles\n\nLLM 0.22,\ \ the annotated release notes - 17th February 2025\nRun LLMs on macOS using llm-mlx\ \ and Apple's MLX framework - 15th February 2025\nURL-addressable Pyodide Python\ \ environments - 13th February 2025\n\n\n \n\n\nThis is Stuff we figured out about\ \ AI in 2023 by Simon Willison, posted on 31st December 2023.\n\nPart of series\ \ LLMs annual review\n\nStuff we figured out about AI in 2023 - Dec. 31, 2023,\ \ 11:59 p.m. \nThings we learned about LLMs in 2024 - Dec. 31, 2024, 6:07 p.m.\ \ \n\n\n\n blogging\n 69" - source_sentence: What incident involving Google Search is mentioned in the context, and what was the nature of the misinformation? sentences: - 'My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment). That same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:' - 'Terminology aside, I remain skeptical as to their utility based, once again, on the challenge of gullibility. LLMs believe anything you tell them. Any systems that attempts to make meaningful decisions on your behalf will run into the same roadblock: how good is a travel agent, or a digital assistant, or even a research tool if it can’t distinguish truth from fiction? Just the other day Google Search was caught serving up an entirely fake description of the non-existant movie “Encanto 2”. It turned out to be summarizing an imagined movie listing from a fan fiction wiki.' - 'On the other hand, as software engineers we are better placed to take advantage of this than anyone else. We’ve all been given weird coding interns—we can use our deep knowledge to prompt them to solve coding problems more effectively than anyone else can. The ethics of this space remain diabolically complex In September last year Andy Baio and I produced the first major story on the unlicensed training data behind Stable Diffusion. Since then, almost every major LLM (and most of the image generation models) have also been trained on unlicensed data.' - source_sentence: What are the limitations of Apple's LLM features compared to frontier LLMs, according to the context? sentences: - 'DeepSeek v3 is a huge 685B parameter model—one of the largest openly licensed models currently available, significantly bigger than the largest of Meta’s Llama series, Llama 3.1 405B. Benchmarks put it up there with Claude 3.5 Sonnet. Vibe benchmarks (aka the Chatbot Arena) currently rank it 7th, just behind the Gemini 2.0 and OpenAI 4o/o1 models. This is by far the highest ranking openly licensed model. The really impressive thing about DeepSeek v3 is the training cost. The model was trained on 2,788,000 H800 GPU hours at an estimated cost of $5,576,000. Llama 3.1 405B trained 30,840,000 GPU hours—11x that used by DeepSeek v3, for a model that benchmarks slightly worse.' - 'An interesting point of comparison here could be the way railways rolled out around the world in the 1800s. Constructing these required enormous investments and had a massive environmental impact, and many of the lines that were built turned out to be unnecessary—sometimes multiple lines from different companies serving the exact same routes! The resulting bubbles contributed to several financial crashes, see Wikipedia for Panic of 1873, Panic of 1893, Panic of 1901 and the UK’s Railway Mania. They left us with a lot of useful infrastructure and a great deal of bankruptcies and environmental damage. The year of slop' - 'Now that those features are rolling out they’re pretty weak. As an LLM power-user I know what these models are capable of, and Apple’s LLM features offer a pale imitation of what a frontier LLM can do. Instead we’re getting notification summaries that misrepresent news headlines and writing assistant tools that I’ve not found useful at all. Genmoji are kind of fun though. The rise of inference-scaling “reasoning” models The most interesting development in the final quarter of 2024 was the introduction of a new shape of LLM, exemplified by OpenAI’s o1 models—initially released as o1-preview and o1-mini on September 12th.' - source_sentence: What new feature was introduced in ChatGPT's voice mode in December? sentences: - The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did. - 'The two main categories I see are people who think AI agents are obviously things that go and act on your behalf—the travel agent model—and people who think in terms of LLMs that have been given access to tools which they can run in a loop as part of solving a problem. The term “autonomy” is often thrown into the mix too, again without including a clear definition. (I also collected 211 definitions on Twitter a few months ago—here they are in Datasette Lite—and had gemini-exp-1206 attempt to summarize them.) Whatever the term may mean, agents still have that feeling of perpetually “coming soon”.' - 'The GPT-4 barrier was comprehensively broken In my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t? I’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.' model-index: - name: SentenceTransformer based on Snowflake/snowflake-arctic-embed-l results: - task: type: information-retrieval name: Information Retrieval dataset: name: Unknown type: unknown metrics: - type: cosine_accuracy@1 value: 0.9166666666666666 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 1.0 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 1.0 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 1.0 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.9166666666666666 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.3333333333333333 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.20000000000000004 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.10000000000000002 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.9166666666666666 name: Cosine Recall@1 - type: cosine_recall@3 value: 1.0 name: Cosine Recall@3 - type: cosine_recall@5 value: 1.0 name: Cosine Recall@5 - type: cosine_recall@10 value: 1.0 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.9692441461309548 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.9583333333333334 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.9583333333333334 name: Cosine Map@100 --- # SentenceTransformer based on Snowflake/snowflake-arctic-embed-l This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l) <!-- at revision d8fb21ca8d905d2832ee8b96c894d3298964346b --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 1024 dimensions - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("dabraldeepti25/legal-ft-v0-updated") # Run inference sentences = [ "What new feature was introduced in ChatGPT's voice mode in December?", 'The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did.', 'The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 1024] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Information Retrieval * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.9167 \| \| cosine_accuracy@3 \| 1.0 \| \| cosine_accuracy@5 \| 1.0 \| \| cosine_accuracy@10 \| 1.0 \| \| cosine_precision@1 \| 0.9167 \| \| cosine_precision@3 \| 0.3333 \| \| cosine_precision@5 \| 0.2 \| \| cosine_precision@10 \| 0.1 \| \| cosine_recall@1 \| 0.9167 \| \| cosine_recall@3 \| 1.0 \| \| cosine_recall@5 \| 1.0 \| \| cosine_recall@10 \| 1.0 \| \| cosine_ndcg@10 \| 0.9692 \| \| cosine_mrr@10 \| 0.9583 \| \| cosine_map@100 \| 0.9583 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 156 training samples * Columns: <code>sentence_0</code> and <code>sentence_1</code> * Approximate statistics based on the first 156 samples: \| \| sentence_0 \| sentence_1 \| \|:--------\|:-----------------------------------------------------------------------------------\|:-------------------------------------------------------------------------------------\| \| type \| string \| string \| \| details \| <ul><li>min: 13 tokens</li><li>mean: 20.17 tokens</li><li>max: 34 tokens</li></ul> \| <ul><li>min: 43 tokens</li><li>mean: 135.18 tokens</li><li>max: 214 tokens</li></ul> \| * Samples: \| sentence_0 \| sentence_1 \| \|:-------------------------------------------------------------------------------------------------------\|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| <code>What is the significance of prompt engineering in DALL-E 3?</code> \| <code>Now add a walrus: Prompt engineering in DALL-E 3<br>32.8k<br>41.2k<br><br><br>Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive<br>32.5k<br>38.2k<br><br><br>ChatGPT can’t access the internet, even though it really looks like it can<br>30.5k<br>34.2k<br><br><br>Stanford Alpaca, and the acceleration of on-device large language model development<br>29.7k<br>35.7k<br><br><br>Run Llama 2 on your own Mac using LLM and Homebrew<br>27.9k<br>33.6k<br><br><br>Midjourney 5.1<br>26.7k<br>33.4k<br><br><br>Think of language models like ChatGPT as a “calculator for words”<br>25k<br>31.8k<br><br><br>Multi-modal prompt injection image attacks against GPT-4V<br>23.7k<br>27.4k</code> \| \| <code>How does the vicuna-7b Large Language Model operate within a browser?</code> \| <code>Now add a walrus: Prompt engineering in DALL-E 3<br>32.8k<br>41.2k<br><br><br>Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive<br>32.5k<br>38.2k<br><br><br>ChatGPT can’t access the internet, even though it really looks like it can<br>30.5k<br>34.2k<br><br><br>Stanford Alpaca, and the acceleration of on-device large language model development<br>29.7k<br>35.7k<br><br><br>Run Llama 2 on your own Mac using LLM and Homebrew<br>27.9k<br>33.6k<br><br><br>Midjourney 5.1<br>26.7k<br>33.4k<br><br><br>Think of language models like ChatGPT as a “calculator for words”<br>25k<br>31.8k<br><br><br>Multi-modal prompt injection image attacks against GPT-4V<br>23.7k<br>27.4k</code> \| \| <code>What model of MacBook Pro is being used in the context, and what is its storage capacity?</code> \| <code>My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment).<br>That same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:</code> \| * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 768, 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `num_train_epochs`: 10 - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `torch_empty_cache_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.0 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: False - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: None - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `include_for_metrics`: [] - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin </details> ### Training Logs \| Epoch \| Step \| cosine_ndcg@10 \| \|:-----:\|:----:\|:--------------:\| \| 1.0 \| 16 \| 0.9692 \| \| 2.0 \| 32 \| 0.9692 \| \| 3.0 \| 48 \| 1.0 \| \| 3.125 \| 50 \| 1.0 \| \| 4.0 \| 64 \| 1.0 \| \| 5.0 \| 80 \| 0.9692 \| \| 6.0 \| 96 \| 0.9692 \| \| 6.25 \| 100 \| 0.9692 \| \| 7.0 \| 112 \| 0.9692 \| \| 8.0 \| 128 \| 0.9692 \| \| 9.0 \| 144 \| 0.9692 \| \| 9.375 \| 150 \| 0.9692 \| \| 10.0 \| 160 \| 0.9692 \| ### Framework Versions - Python: 3.11.11 - Sentence Transformers: 3.4.1 - Transformers: 4.48.3 - PyTorch: 2.5.1+cu124 - Accelerate: 1.3.0 - Datasets: 3.3.1 - Tokenizers: 0.21.0 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# SentenceTransformer based on Snowflake/snowflake-arctic-embed-l This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l) <!-- at revision d8fb21ca8d905d2832ee8b96c894d3298964346b --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 1024 dimensions - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("dabraldeepti25/legal-ft-v0-updated") # Run inference sentences = [ "What new feature was introduced in ChatGPT's voice mode in December?", 'The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did.', 'The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 1024] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Information Retrieval * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.9167 \| \| cosine_accuracy@3 \| 1.0 \| \| cosine_accuracy@5 \| 1.0 \| \| cosine_accuracy@10 \| 1.0 \| \| cosine_precision@1 \| 0.9167 \| \| cosine_precision@3 \| 0.3333 \| \| cosine_precision@5 \| 0.2 \| \| cosine_precision@10 \| 0.1 \| \| cosine_recall@1 \| 0.9167 \| \| cosine_recall@3 \| 1.0 \| \| cosine_recall@5 \| 1.0 \| \| cosine_recall@10 \| 1.0 \| \| cosine_ndcg@10 \| 0.9692 \| \| cosine_mrr@10 \| 0.9583 \| \| cosine_map@100 \| 0.9583 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 156 training samples * Columns: <code>sentence_0</code> and <code>sentence_1</code> * Approximate statistics based on the first 156 samples: \| \| sentence_0 \| sentence_1 \| \|:--------\|:-----------------------------------------------------------------------------------\|:-------------------------------------------------------------------------------------\| \| type \| string \| string \| \| details \| <ul><li>min: 13 tokens</li><li>mean: 20.17 tokens</li><li>max: 34 tokens</li></ul> \| <ul><li>min: 43 tokens</li><li>mean: 135.18 tokens</li><li>max: 214 tokens</li></ul> \| * Samples: \| sentence_0 \| sentence_1 \| \|:-------------------------------------------------------------------------------------------------------\|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| <code>What is the significance of prompt engineering in DALL-E 3?</code> \| <code>Now add a walrus: Prompt engineering in DALL-E 3<br>32.8k<br>41.2k<br><br><br>Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive<br>32.5k<br>38.2k<br><br><br>ChatGPT can’t access the internet, even though it really looks like it can<br>30.5k<br>34.2k<br><br><br>Stanford Alpaca, and the acceleration of on-device large language model development<br>29.7k<br>35.7k<br><br><br>Run Llama 2 on your own Mac using LLM and Homebrew<br>27.9k<br>33.6k<br><br><br>Midjourney 5.1<br>26.7k<br>33.4k<br><br><br>Think of language models like ChatGPT as a “calculator for words”<br>25k<br>31.8k<br><br><br>Multi-modal prompt injection image attacks against GPT-4V<br>23.7k<br>27.4k</code> \| \| <code>How does the vicuna-7b Large Language Model operate within a browser?</code> \| <code>Now add a walrus: Prompt engineering in DALL-E 3<br>32.8k<br>41.2k<br><br><br>Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive<br>32.5k<br>38.2k<br><br><br>ChatGPT can’t access the internet, even though it really looks like it can<br>30.5k<br>34.2k<br><br><br>Stanford Alpaca, and the acceleration of on-device large language model development<br>29.7k<br>35.7k<br><br><br>Run Llama 2 on your own Mac using LLM and Homebrew<br>27.9k<br>33.6k<br><br><br>Midjourney 5.1<br>26.7k<br>33.4k<br><br><br>Think of language models like ChatGPT as a “calculator for words”<br>25k<br>31.8k<br><br><br>Multi-modal prompt injection image attacks against GPT-4V<br>23.7k<br>27.4k</code> \| \| <code>What model of MacBook Pro is being used in the context, and what is its storage capacity?</code> \| <code>My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment).<br>That same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:</code> \| * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 768, 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `num_train_epochs`: 10 - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `torch_empty_cache_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.0 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: False - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: None - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `include_for_metrics`: [] - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin </details> ### Training Logs \| Epoch \| Step \| cosine_ndcg@10 \| \|:-----:\|:----:\|:--------------:\| \| 1.0 \| 16 \| 0.9692 \| \| 2.0 \| 32 \| 0.9692 \| \| 3.0 \| 48 \| 1.0 \| \| 3.125 \| 50 \| 1.0 \| \| 4.0 \| 64 \| 1.0 \| \| 5.0 \| 80 \| 0.9692 \| \| 6.0 \| 96 \| 0.9692 \| \| 6.25 \| 100 \| 0.9692 \| \| 7.0 \| 112 \| 0.9692 \| \| 8.0 \| 128 \| 0.9692 \| \| 9.0 \| 144 \| 0.9692 \| \| 9.375 \| 150 \| 0.9692 \| \| 10.0 \| 160 \| 0.9692 \| ### Framework Versions - Python: 3.11.11 - Sentence Transformers: 3.4.1 - Transformers: 4.48.3 - PyTorch: 2.5.1+cu124 - Accelerate: 1.3.0 - Datasets: 3.3.1 - Tokenizers: 0.21.0 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "Snowflake/snowflake-arctic-embed-l", "library_name": "sentence-transformers", "metrics": ["cosine_accuracy@1", "cosine_accuracy@3", "cosine_accuracy@5", "cosine_accuracy@10", "cosine_precision@1", "cosine_precision@3", "cosine_precision@5", "cosine_precision@10", "cosine_recall@1", "cosine_recall@3", "cosine_recall@5", "cosine_recall@10", "cosine_ndcg@10", "cosine_mrr@10", "cosine_map@100"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:156", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss"], "widget": [{"source_sentence": "What significant change occurred in the AI landscape regarding models surpassing GPT-4 in the past twelve months?", "sentences": ["Except... you can run generated code to see if it’s correct. And with patterns like ChatGPT Code Interpreter the LLM can execute the code itself, process the error message, then rewrite it and keep trying until it works!\nSo hallucination is a much lesser problem for code generation than for anything else. If only we had the equivalent of Code Interpreter for fact-checking natural language!\nHow should we feel about this as software engineers?\nOn the one hand, this feels like a threat: who needs a programmer if ChatGPT can write code for you?", "The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.", "If you think about what they do, this isn’t such a big surprise. The grammar rules of programming languages like Python and JavaScript are massively less complicated than the grammar of Chinese, Spanish or English.\nIt’s still astonishing to me how effective they are though.\nOne of the great weaknesses of LLMs is their tendency to hallucinate—to imagine things that don’t correspond to reality. You would expect this to be a particularly bad problem for code—if an LLM hallucinates a method that doesn’t exist, the code should be useless."]}, {"source_sentence": "How does Claude enable users to interact with applications created through its interface?", "sentences": ["OpenAI made GPT-4o free for all users in May, and Claude 3.5 Sonnet was freely available from its launch in June. This was a momentus change, because for the previous year free users had mostly been restricted to GPT-3.5 level models, meaning new users got a very inaccurate mental model of what a capable LLM could actually do.\nThat era appears to have ended, likely permanently, with OpenAI’s launch of ChatGPT Pro. This $200/month subscription service is the only way to access their most capable model, o1 Pro.\nSince the trick behind the o1 series (and the future models it will undoubtedly inspire) is to expend more compute time to get better results, I don’t think those days of free access to the best available models are likely to return.", "We already knew LLMs were spookily good at writing code. If you prompt them right, it turns out they can build you a full interactive application using HTML, CSS and JavaScript (and tools like React if you wire up some extra supporting build mechanisms)—often in a single prompt.\nAnthropic kicked this idea into high gear when they released Claude Artifacts, a groundbreaking new feature that was initially slightly lost in the noise due to being described half way through their announcement of the incredible Claude 3.5 Sonnet.\nWith Artifacts, Claude can write you an on-demand interactive application and then let you use it directly inside the Claude interface.\nHere’s my Extract URLs app, entirely generated by Claude:", "Industry’s Tardy Response to the AI Prompt Injection Vulnerability on RedMonk Conversations\n\n\nPosted 31st December 2023 at 11:59 pm · Follow me on Mastodon or Twitter or subscribe to my newsletter\n\n\nMore recent articles\n\nLLM 0.22, the annotated release notes - 17th February 2025\nRun LLMs on macOS using llm-mlx and Apple's MLX framework - 15th February 2025\nURL-addressable Pyodide Python environments - 13th February 2025\n\n\n \n\n\nThis is Stuff we figured out about AI in 2023 by Simon Willison, posted on 31st December 2023.\n\nPart of series LLMs annual review\n\nStuff we figured out about AI in 2023 - Dec. 31, 2023, 11:59 p.m. \nThings we learned about LLMs in 2024 - Dec. 31, 2024, 6:07 p.m. \n\n\n\n blogging\n 69"]}, {"source_sentence": "What incident involving Google Search is mentioned in the context, and what was the nature of the misinformation?", "sentences": ["My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment).\nThat same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:", "Terminology aside, I remain skeptical as to their utility based, once again, on the challenge of gullibility. LLMs believe anything you tell them. Any systems that attempts to make meaningful decisions on your behalf will run into the same roadblock: how good is a travel agent, or a digital assistant, or even a research tool if it can’t distinguish truth from fiction?\nJust the other day Google Search was caught serving up an entirely fake description of the non-existant movie “Encanto 2”. It turned out to be summarizing an imagined movie listing from a fan fiction wiki.", "On the other hand, as software engineers we are better placed to take advantage of this than anyone else. We’ve all been given weird coding interns—we can use our deep knowledge to prompt them to solve coding problems more effectively than anyone else can.\nThe ethics of this space remain diabolically complex\nIn September last year Andy Baio and I produced the first major story on the unlicensed training data behind Stable Diffusion.\nSince then, almost every major LLM (and most of the image generation models) have also been trained on unlicensed data."]}, {"source_sentence": "What are the limitations of Apple's LLM features compared to frontier LLMs, according to the context?", "sentences": ["DeepSeek v3 is a huge 685B parameter model—one of the largest openly licensed models currently available, significantly bigger than the largest of Meta’s Llama series, Llama 3.1 405B.\nBenchmarks put it up there with Claude 3.5 Sonnet. Vibe benchmarks (aka the Chatbot Arena) currently rank it 7th, just behind the Gemini 2.0 and OpenAI 4o/o1 models. This is by far the highest ranking openly licensed model.\nThe really impressive thing about DeepSeek v3 is the training cost. The model was trained on 2,788,000 H800 GPU hours at an estimated cost of $5,576,000. Llama 3.1 405B trained 30,840,000 GPU hours—11x that used by DeepSeek v3, for a model that benchmarks slightly worse.", "An interesting point of comparison here could be the way railways rolled out around the world in the 1800s. Constructing these required enormous investments and had a massive environmental impact, and many of the lines that were built turned out to be unnecessary—sometimes multiple lines from different companies serving the exact same routes!\nThe resulting bubbles contributed to several financial crashes, see Wikipedia for Panic of 1873, Panic of 1893, Panic of 1901 and the UK’s Railway Mania. They left us with a lot of useful infrastructure and a great deal of bankruptcies and environmental damage.\nThe year of slop", "Now that those features are rolling out they’re pretty weak. As an LLM power-user I know what these models are capable of, and Apple’s LLM features offer a pale imitation of what a frontier LLM can do. Instead we’re getting notification summaries that misrepresent news headlines and writing assistant tools that I’ve not found useful at all. Genmoji are kind of fun though.\nThe rise of inference-scaling “reasoning” models\nThe most interesting development in the final quarter of 2024 was the introduction of a new shape of LLM, exemplified by OpenAI’s o1 models—initially released as o1-preview and o1-mini on September 12th."]}, {"source_sentence": "What new feature was introduced in ChatGPT's voice mode in December?", "sentences": ["The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did.", "The two main categories I see are people who think AI agents are obviously things that go and act on your behalf—the travel agent model—and people who think in terms of LLMs that have been given access to tools which they can run in a loop as part of solving a problem. The term “autonomy” is often thrown into the mix too, again without including a clear definition.\n(I also collected 211 definitions on Twitter a few months ago—here they are in Datasette Lite—and had gemini-exp-1206 attempt to summarize them.)\nWhatever the term may mean, agents still have that feeling of perpetually “coming soon”.", "The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total."]}], "model-index": [{"name": "SentenceTransformer based on Snowflake/snowflake-arctic-embed-l", "results": [{"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "Unknown", "type": "unknown"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.9166666666666666, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 1.0, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 1.0, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 1.0, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.9166666666666666, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.3333333333333333, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.20000000000000004, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.10000000000000002, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.9166666666666666, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 1.0, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 1.0, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 1.0, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.9692441461309548, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.9583333333333334, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.9583333333333334, "name": "Cosine Map@100"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,533
MoritzLaurer/deberta-v3-base-zeroshot-v2.0-c	MoritzLaurer	zero-shot-classification	[ "transformers", "safetensors", "deberta-v2", "text-classification", "zero-shot-classification", "en", "arxiv:2312.17543", "base_model:microsoft/deberta-v3-base", "base_model:finetune:microsoft/deberta-v3-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2024-03-21T22:21:55Z	2024-04-04T07:04:21+00:00	369	0	--- base_model: microsoft/deberta-v3-base language: - en library_name: transformers license: mit pipeline_tag: zero-shot-classification tags: - text-classification - zero-shot-classification --- # Model description: deberta-v3-base-zeroshot-v2.0-c ## zeroshot-v2.0 series of models Models in this series are designed for efficient zeroshot classification with the Hugging Face pipeline. These models can do classification without training data and run on both GPUs and CPUs. An overview of the latest zeroshot classifiers is available in my [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). The main update of this `zeroshot-v2.0` series of models is that several models are trained on fully commercially-friendly data for users with strict license requirements. These models can do one universal classification task: determine whether a hypothesis is "true" or "not true" given a text (`entailment` vs. `not_entailment`). This task format is based on the Natural Language Inference task (NLI). The task is so universal that any classification task can be reformulated into this task by the Hugging Face pipeline. ## Training data Models with a "`-c`" in the name are trained on two types of fully commercially-friendly data: 1. Synthetic data generated with [Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1). I first created a list of 500+ diverse text classification tasks for 25 professions in conversations with Mistral-large. The data was manually curated. I then used this as seed data to generate several hundred thousand texts for these tasks with Mixtral-8x7B-Instruct-v0.1. The final dataset used is available in the [synthetic_zeroshot_mixtral_v0.1](https://huggingface.co/datasets/MoritzLaurer/synthetic_zeroshot_mixtral_v0.1) dataset in the subset `mixtral_written_text_for_tasks_v4`. Data curation was done in multiple iterations and will be improved in future iterations. 2. Two commercially-friendly NLI datasets: ([MNLI](https://huggingface.co/datasets/nyu-mll/multi_nli), [FEVER-NLI](https://huggingface.co/datasets/fever)). These datasets were added to increase generalization. 3. Models without a "`-c`" in the name also included a broader mix of training data with a broader mix of licenses: ANLI, WANLI, LingNLI, and all datasets in [this list](https://github.com/MoritzLaurer/zeroshot-classifier/blob/7f82e4ab88d7aa82a4776f161b368cc9fa778001/v1_human_data/datasets_overview.csv) where `used_in_v1.1==True`. ## How to use the models ```python #!pip install transformers[sentencepiece] from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ``` `multi_label=False` forces the model to decide on only one class. `multi_label=True` enables the model to choose multiple classes. ## Metrics The models were evaluated on 28 different text classification tasks with the [f1_macro](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) metric. The main reference point is `facebook/bart-large-mnli` which is, at the time of writing (03.04.24), the most used commercially-friendly 0-shot classifier. ![results_aggreg_v2.0](https://raw.githubusercontent.com/MoritzLaurer/zeroshot-classifier/main/v2_synthetic_data/results/zeroshot-v2.0-aggreg.png) \| \| facebook/bart-large-mnli \| roberta-base-zeroshot-v2.0-c \| roberta-large-zeroshot-v2.0-c \| deberta-v3-base-zeroshot-v2.0-c \| deberta-v3-base-zeroshot-v2.0 (fewshot) \| deberta-v3-large-zeroshot-v2.0-c \| deberta-v3-large-zeroshot-v2.0 (fewshot) \| bge-m3-zeroshot-v2.0-c \| bge-m3-zeroshot-v2.0 (fewshot) \| \|:---------------------------\|---------------------------:\|-----------------------------:\|------------------------------:\|--------------------------------:\|-----------------------------------:\|---------------------------------:\|------------------------------------:\|-----------------------:\|--------------------------:\| \| all datasets mean \| 0.497 \| 0.587 \| 0.622 \| 0.619 \| 0.643 (0.834) \| 0.676 \| 0.673 (0.846) \| 0.59 \| (0.803) \| \| amazonpolarity (2) \| 0.937 \| 0.924 \| 0.951 \| 0.937 \| 0.943 (0.961) \| 0.952 \| 0.956 (0.968) \| 0.942 \| (0.951) \| \| imdb (2) \| 0.892 \| 0.871 \| 0.904 \| 0.893 \| 0.899 (0.936) \| 0.923 \| 0.918 (0.958) \| 0.873 \| (0.917) \| \| appreviews (2) \| 0.934 \| 0.913 \| 0.937 \| 0.938 \| 0.945 (0.948) \| 0.943 \| 0.949 (0.962) \| 0.932 \| (0.954) \| \| yelpreviews (2) \| 0.948 \| 0.953 \| 0.977 \| 0.979 \| 0.975 (0.989) \| 0.988 \| 0.985 (0.994) \| 0.973 \| (0.978) \| \| rottentomatoes (2) \| 0.83 \| 0.802 \| 0.841 \| 0.84 \| 0.86 (0.902) \| 0.869 \| 0.868 (0.908) \| 0.813 \| (0.866) \| \| emotiondair (6) \| 0.455 \| 0.482 \| 0.486 \| 0.459 \| 0.495 (0.748) \| 0.499 \| 0.484 (0.688) \| 0.453 \| (0.697) \| \| emocontext (4) \| 0.497 \| 0.555 \| 0.63 \| 0.59 \| 0.592 (0.799) \| 0.699 \| 0.676 (0.81) \| 0.61 \| (0.798) \| \| empathetic (32) \| 0.371 \| 0.374 \| 0.404 \| 0.378 \| 0.405 (0.53) \| 0.447 \| 0.478 (0.555) \| 0.387 \| (0.455) \| \| financialphrasebank (3) \| 0.465 \| 0.562 \| 0.455 \| 0.714 \| 0.669 (0.906) \| 0.691 \| 0.582 (0.913) \| 0.504 \| (0.895) \| \| banking77 (72) \| 0.312 \| 0.124 \| 0.29 \| 0.421 \| 0.446 (0.751) \| 0.513 \| 0.567 (0.766) \| 0.387 \| (0.715) \| \| massive (59) \| 0.43 \| 0.428 \| 0.543 \| 0.512 \| 0.52 (0.755) \| 0.526 \| 0.518 (0.789) \| 0.414 \| (0.692) \| \| wikitoxic_toxicaggreg (2) \| 0.547 \| 0.751 \| 0.766 \| 0.751 \| 0.769 (0.904) \| 0.741 \| 0.787 (0.911) \| 0.736 \| (0.9) \| \| wikitoxic_obscene (2) \| 0.713 \| 0.817 \| 0.854 \| 0.853 \| 0.869 (0.922) \| 0.883 \| 0.893 (0.933) \| 0.783 \| (0.914) \| \| wikitoxic_threat (2) \| 0.295 \| 0.71 \| 0.817 \| 0.813 \| 0.87 (0.946) \| 0.827 \| 0.879 (0.952) \| 0.68 \| (0.947) \| \| wikitoxic_insult (2) \| 0.372 \| 0.724 \| 0.798 \| 0.759 \| 0.811 (0.912) \| 0.77 \| 0.779 (0.924) \| 0.783 \| (0.915) \| \| wikitoxic_identityhate (2) \| 0.473 \| 0.774 \| 0.798 \| 0.774 \| 0.765 (0.938) \| 0.797 \| 0.806 (0.948) \| 0.761 \| (0.931) \| \| hateoffensive (3) \| 0.161 \| 0.352 \| 0.29 \| 0.315 \| 0.371 (0.862) \| 0.47 \| 0.461 (0.847) \| 0.291 \| (0.823) \| \| hatexplain (3) \| 0.239 \| 0.396 \| 0.314 \| 0.376 \| 0.369 (0.765) \| 0.378 \| 0.389 (0.764) \| 0.29 \| (0.729) \| \| biasframes_offensive (2) \| 0.336 \| 0.571 \| 0.583 \| 0.544 \| 0.601 (0.867) \| 0.644 \| 0.656 (0.883) \| 0.541 \| (0.855) \| \| biasframes_sex (2) \| 0.263 \| 0.617 \| 0.835 \| 0.741 \| 0.809 (0.922) \| 0.846 \| 0.815 (0.946) \| 0.748 \| (0.905) \| \| biasframes_intent (2) \| 0.616 \| 0.531 \| 0.635 \| 0.554 \| 0.61 (0.881) \| 0.696 \| 0.687 (0.891) \| 0.467 \| (0.868) \| \| agnews (4) \| 0.703 \| 0.758 \| 0.745 \| 0.68 \| 0.742 (0.898) \| 0.819 \| 0.771 (0.898) \| 0.687 \| (0.892) \| \| yahootopics (10) \| 0.299 \| 0.543 \| 0.62 \| 0.578 \| 0.564 (0.722) \| 0.621 \| 0.613 (0.738) \| 0.587 \| (0.711) \| \| trueteacher (2) \| 0.491 \| 0.469 \| 0.402 \| 0.431 \| 0.479 (0.82) \| 0.459 \| 0.538 (0.846) \| 0.471 \| (0.518) \| \| spam (2) \| 0.505 \| 0.528 \| 0.504 \| 0.507 \| 0.464 (0.973) \| 0.74 \| 0.597 (0.983) \| 0.441 \| (0.978) \| \| wellformedquery (2) \| 0.407 \| 0.333 \| 0.333 \| 0.335 \| 0.491 (0.769) \| 0.334 \| 0.429 (0.815) \| 0.361 \| (0.718) \| \| manifesto (56) \| 0.084 \| 0.102 \| 0.182 \| 0.17 \| 0.187 (0.376) \| 0.258 \| 0.256 (0.408) \| 0.147 \| (0.331) \| \| capsotu (21) \| 0.34 \| 0.479 \| 0.523 \| 0.502 \| 0.477 (0.664) \| 0.603 \| 0.502 (0.686) \| 0.472 \| (0.644) \| These numbers indicate zeroshot performance, as no data from these datasets was added in the training mix. Note that models without a "`-c`" in the title were evaluated twice: one run without any data from these 28 datasets to test pure zeroshot performance (the first number in the respective column) and the final run including up to 500 training data points per class from each of the 28 datasets (the second number in brackets in the column, "fewshot"). No model was trained on test data. Details on the different datasets are available here: https://github.com/MoritzLaurer/zeroshot-classifier/blob/main/v1_human_data/datasets_overview.csv ## When to use which model - deberta-v3-zeroshot vs. roberta-zeroshot: deberta-v3 performs clearly better than roberta, but it is a bit slower. roberta is directly compatible with Hugging Face's production inference TEI containers and flash attention. These containers are a good choice for production use-cases. tl;dr: For accuracy, use a deberta-v3 model. If production inference speed is a concern, you can consider a roberta model (e.g. in a TEI container and [HF Inference Endpoints](https://ui.endpoints.huggingface.co/catalog)). - commercial use-cases: models with "`-c`" in the title are guaranteed to be trained on only commercially-friendly data. Models without a "`-c`" were trained on more data and perform better, but include data with non-commercial licenses. Legal opinions diverge if this training data affects the license of the trained model. For users with strict legal requirements, the models with "`-c`" in the title are recommended. - Multilingual/non-English use-cases: use [bge-m3-zeroshot-v2.0](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0) or [bge-m3-zeroshot-v2.0-c](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0-c). Note that multilingual models perform worse than English-only models. You can therefore also first machine translate your texts to English with libraries like [EasyNMT](https://github.com/UKPLab/EasyNMT) and then apply any English-only model to the translated data. Machine translation also facilitates validation in case your team does not speak all languages in the data. - context window: The `bge-m3` models can process up to 8192 tokens. The other models can process up to 512. Note that longer text inputs both make the mode slower and decrease performance, so if you're only working with texts of up to 400~ words / 1 page, use e.g. a deberta model for better performance. - The latest updates on new models are always available in the [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). ## Reproduction Reproduction code is available in the `v2_synthetic_data` directory here: https://github.com/MoritzLaurer/zeroshot-classifier/tree/main ## Limitations and bias The model can only do text classification tasks. Biases can come from the underlying foundation model, the human NLI training data and the synthetic data generated by Mixtral. ## License The foundation model was published under the MIT license. The licenses of the training data vary depending on the model, see above. ## Citation This model is an extension of the research described in this [paper](https://arxiv.org/pdf/2312.17543.pdf). If you use this model academically, please cite: ``` @misc{laurer_building_2023, title = {Building {Efficient} {Universal} {Classifiers} with {Natural} {Language} {Inference}}, url = {http://arxiv.org/abs/2312.17543}, doi = {10.48550/arXiv.2312.17543}, abstract = {Generative Large Language Models (LLMs) have become the mainstream choice for fewshot and zeroshot learning thanks to the universality of text generation. Many users, however, do not need the broad capabilities of generative LLMs when they only want to automate a classification task. Smaller BERT-like models can also learn universal tasks, which allow them to do any text classification task without requiring fine-tuning (zeroshot classification) or to learn new tasks with only a few examples (fewshot), while being significantly more efficient than generative LLMs. This paper (1) explains how Natural Language Inference (NLI) can be used as a universal classification task that follows similar principles as instruction fine-tuning of generative LLMs, (2) provides a step-by-step guide with reusable Jupyter notebooks for building a universal classifier, and (3) shares the resulting universal classifier that is trained on 33 datasets with 389 diverse classes. Parts of the code we share has been used to train our older zeroshot classifiers that have been downloaded more than 55 million times via the Hugging Face Hub as of December 2023. Our new classifier improves zeroshot performance by 9.4\%.}, urldate = {2024-01-05}, publisher = {arXiv}, author = {Laurer, Moritz and van Atteveldt, Wouter and Casas, Andreu and Welbers, Kasper}, month = dec, year = {2023}, note = {arXiv:2312.17543 [cs]}, keywords = {Computer Science - Artificial Intelligence, Computer Science - Computation and Language}, } ``` ### Ideas for cooperation or questions? If you have questions or ideas for cooperation, contact me at moritz{at}huggingface{dot}co or [LinkedIn](https://www.linkedin.com/in/moritz-laurer/) ### Flexible usage and "prompting" You can formulate your own hypotheses by changing the `hypothesis_template` of the zeroshot pipeline. Similar to "prompt engineering" for LLMs, you can test different formulations of your `hypothesis_template` and verbalized classes to improve performance. ```python from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" # formulation 1 hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] # formulation 2 depending on your use-case hypothesis_template = "The topic of this text is {}" classes_verbalized = ["political activities", "economic policy", "entertainment or music", "environmental protection"] # test different formulations zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ```	null	Non_BioNLP	# Model description: deberta-v3-base-zeroshot-v2.0-c ## zeroshot-v2.0 series of models Models in this series are designed for efficient zeroshot classification with the Hugging Face pipeline. These models can do classification without training data and run on both GPUs and CPUs. An overview of the latest zeroshot classifiers is available in my [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). The main update of this `zeroshot-v2.0` series of models is that several models are trained on fully commercially-friendly data for users with strict license requirements. These models can do one universal classification task: determine whether a hypothesis is "true" or "not true" given a text (`entailment` vs. `not_entailment`). This task format is based on the Natural Language Inference task (NLI). The task is so universal that any classification task can be reformulated into this task by the Hugging Face pipeline. ## Training data Models with a "`-c`" in the name are trained on two types of fully commercially-friendly data: 1. Synthetic data generated with [Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1). I first created a list of 500+ diverse text classification tasks for 25 professions in conversations with Mistral-large. The data was manually curated. I then used this as seed data to generate several hundred thousand texts for these tasks with Mixtral-8x7B-Instruct-v0.1. The final dataset used is available in the [synthetic_zeroshot_mixtral_v0.1](https://huggingface.co/datasets/MoritzLaurer/synthetic_zeroshot_mixtral_v0.1) dataset in the subset `mixtral_written_text_for_tasks_v4`. Data curation was done in multiple iterations and will be improved in future iterations. 2. Two commercially-friendly NLI datasets: ([MNLI](https://huggingface.co/datasets/nyu-mll/multi_nli), [FEVER-NLI](https://huggingface.co/datasets/fever)). These datasets were added to increase generalization. 3. Models without a "`-c`" in the name also included a broader mix of training data with a broader mix of licenses: ANLI, WANLI, LingNLI, and all datasets in [this list](https://github.com/MoritzLaurer/zeroshot-classifier/blob/7f82e4ab88d7aa82a4776f161b368cc9fa778001/v1_human_data/datasets_overview.csv) where `used_in_v1.1==True`. ## How to use the models ```python #!pip install transformers[sentencepiece] from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ``` `multi_label=False` forces the model to decide on only one class. `multi_label=True` enables the model to choose multiple classes. ## Metrics The models were evaluated on 28 different text classification tasks with the [f1_macro](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) metric. The main reference point is `facebook/bart-large-mnli` which is, at the time of writing (03.04.24), the most used commercially-friendly 0-shot classifier. ![results_aggreg_v2.0](https://raw.githubusercontent.com/MoritzLaurer/zeroshot-classifier/main/v2_synthetic_data/results/zeroshot-v2.0-aggreg.png) \| \| facebook/bart-large-mnli \| roberta-base-zeroshot-v2.0-c \| roberta-large-zeroshot-v2.0-c \| deberta-v3-base-zeroshot-v2.0-c \| deberta-v3-base-zeroshot-v2.0 (fewshot) \| deberta-v3-large-zeroshot-v2.0-c \| deberta-v3-large-zeroshot-v2.0 (fewshot) \| bge-m3-zeroshot-v2.0-c \| bge-m3-zeroshot-v2.0 (fewshot) \| \|:---------------------------\|---------------------------:\|-----------------------------:\|------------------------------:\|--------------------------------:\|-----------------------------------:\|---------------------------------:\|------------------------------------:\|-----------------------:\|--------------------------:\| \| all datasets mean \| 0.497 \| 0.587 \| 0.622 \| 0.619 \| 0.643 (0.834) \| 0.676 \| 0.673 (0.846) \| 0.59 \| (0.803) \| \| amazonpolarity (2) \| 0.937 \| 0.924 \| 0.951 \| 0.937 \| 0.943 (0.961) \| 0.952 \| 0.956 (0.968) \| 0.942 \| (0.951) \| \| imdb (2) \| 0.892 \| 0.871 \| 0.904 \| 0.893 \| 0.899 (0.936) \| 0.923 \| 0.918 (0.958) \| 0.873 \| (0.917) \| \| appreviews (2) \| 0.934 \| 0.913 \| 0.937 \| 0.938 \| 0.945 (0.948) \| 0.943 \| 0.949 (0.962) \| 0.932 \| (0.954) \| \| yelpreviews (2) \| 0.948 \| 0.953 \| 0.977 \| 0.979 \| 0.975 (0.989) \| 0.988 \| 0.985 (0.994) \| 0.973 \| (0.978) \| \| rottentomatoes (2) \| 0.83 \| 0.802 \| 0.841 \| 0.84 \| 0.86 (0.902) \| 0.869 \| 0.868 (0.908) \| 0.813 \| (0.866) \| \| emotiondair (6) \| 0.455 \| 0.482 \| 0.486 \| 0.459 \| 0.495 (0.748) \| 0.499 \| 0.484 (0.688) \| 0.453 \| (0.697) \| \| emocontext (4) \| 0.497 \| 0.555 \| 0.63 \| 0.59 \| 0.592 (0.799) \| 0.699 \| 0.676 (0.81) \| 0.61 \| (0.798) \| \| empathetic (32) \| 0.371 \| 0.374 \| 0.404 \| 0.378 \| 0.405 (0.53) \| 0.447 \| 0.478 (0.555) \| 0.387 \| (0.455) \| \| financialphrasebank (3) \| 0.465 \| 0.562 \| 0.455 \| 0.714 \| 0.669 (0.906) \| 0.691 \| 0.582 (0.913) \| 0.504 \| (0.895) \| \| banking77 (72) \| 0.312 \| 0.124 \| 0.29 \| 0.421 \| 0.446 (0.751) \| 0.513 \| 0.567 (0.766) \| 0.387 \| (0.715) \| \| massive (59) \| 0.43 \| 0.428 \| 0.543 \| 0.512 \| 0.52 (0.755) \| 0.526 \| 0.518 (0.789) \| 0.414 \| (0.692) \| \| wikitoxic_toxicaggreg (2) \| 0.547 \| 0.751 \| 0.766 \| 0.751 \| 0.769 (0.904) \| 0.741 \| 0.787 (0.911) \| 0.736 \| (0.9) \| \| wikitoxic_obscene (2) \| 0.713 \| 0.817 \| 0.854 \| 0.853 \| 0.869 (0.922) \| 0.883 \| 0.893 (0.933) \| 0.783 \| (0.914) \| \| wikitoxic_threat (2) \| 0.295 \| 0.71 \| 0.817 \| 0.813 \| 0.87 (0.946) \| 0.827 \| 0.879 (0.952) \| 0.68 \| (0.947) \| \| wikitoxic_insult (2) \| 0.372 \| 0.724 \| 0.798 \| 0.759 \| 0.811 (0.912) \| 0.77 \| 0.779 (0.924) \| 0.783 \| (0.915) \| \| wikitoxic_identityhate (2) \| 0.473 \| 0.774 \| 0.798 \| 0.774 \| 0.765 (0.938) \| 0.797 \| 0.806 (0.948) \| 0.761 \| (0.931) \| \| hateoffensive (3) \| 0.161 \| 0.352 \| 0.29 \| 0.315 \| 0.371 (0.862) \| 0.47 \| 0.461 (0.847) \| 0.291 \| (0.823) \| \| hatexplain (3) \| 0.239 \| 0.396 \| 0.314 \| 0.376 \| 0.369 (0.765) \| 0.378 \| 0.389 (0.764) \| 0.29 \| (0.729) \| \| biasframes_offensive (2) \| 0.336 \| 0.571 \| 0.583 \| 0.544 \| 0.601 (0.867) \| 0.644 \| 0.656 (0.883) \| 0.541 \| (0.855) \| \| biasframes_sex (2) \| 0.263 \| 0.617 \| 0.835 \| 0.741 \| 0.809 (0.922) \| 0.846 \| 0.815 (0.946) \| 0.748 \| (0.905) \| \| biasframes_intent (2) \| 0.616 \| 0.531 \| 0.635 \| 0.554 \| 0.61 (0.881) \| 0.696 \| 0.687 (0.891) \| 0.467 \| (0.868) \| \| agnews (4) \| 0.703 \| 0.758 \| 0.745 \| 0.68 \| 0.742 (0.898) \| 0.819 \| 0.771 (0.898) \| 0.687 \| (0.892) \| \| yahootopics (10) \| 0.299 \| 0.543 \| 0.62 \| 0.578 \| 0.564 (0.722) \| 0.621 \| 0.613 (0.738) \| 0.587 \| (0.711) \| \| trueteacher (2) \| 0.491 \| 0.469 \| 0.402 \| 0.431 \| 0.479 (0.82) \| 0.459 \| 0.538 (0.846) \| 0.471 \| (0.518) \| \| spam (2) \| 0.505 \| 0.528 \| 0.504 \| 0.507 \| 0.464 (0.973) \| 0.74 \| 0.597 (0.983) \| 0.441 \| (0.978) \| \| wellformedquery (2) \| 0.407 \| 0.333 \| 0.333 \| 0.335 \| 0.491 (0.769) \| 0.334 \| 0.429 (0.815) \| 0.361 \| (0.718) \| \| manifesto (56) \| 0.084 \| 0.102 \| 0.182 \| 0.17 \| 0.187 (0.376) \| 0.258 \| 0.256 (0.408) \| 0.147 \| (0.331) \| \| capsotu (21) \| 0.34 \| 0.479 \| 0.523 \| 0.502 \| 0.477 (0.664) \| 0.603 \| 0.502 (0.686) \| 0.472 \| (0.644) \| These numbers indicate zeroshot performance, as no data from these datasets was added in the training mix. Note that models without a "`-c`" in the title were evaluated twice: one run without any data from these 28 datasets to test pure zeroshot performance (the first number in the respective column) and the final run including up to 500 training data points per class from each of the 28 datasets (the second number in brackets in the column, "fewshot"). No model was trained on test data. Details on the different datasets are available here: https://github.com/MoritzLaurer/zeroshot-classifier/blob/main/v1_human_data/datasets_overview.csv ## When to use which model - deberta-v3-zeroshot vs. roberta-zeroshot: deberta-v3 performs clearly better than roberta, but it is a bit slower. roberta is directly compatible with Hugging Face's production inference TEI containers and flash attention. These containers are a good choice for production use-cases. tl;dr: For accuracy, use a deberta-v3 model. If production inference speed is a concern, you can consider a roberta model (e.g. in a TEI container and [HF Inference Endpoints](https://ui.endpoints.huggingface.co/catalog)). - commercial use-cases: models with "`-c`" in the title are guaranteed to be trained on only commercially-friendly data. Models without a "`-c`" were trained on more data and perform better, but include data with non-commercial licenses. Legal opinions diverge if this training data affects the license of the trained model. For users with strict legal requirements, the models with "`-c`" in the title are recommended. - Multilingual/non-English use-cases: use [bge-m3-zeroshot-v2.0](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0) or [bge-m3-zeroshot-v2.0-c](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0-c). Note that multilingual models perform worse than English-only models. You can therefore also first machine translate your texts to English with libraries like [EasyNMT](https://github.com/UKPLab/EasyNMT) and then apply any English-only model to the translated data. Machine translation also facilitates validation in case your team does not speak all languages in the data. - context window: The `bge-m3` models can process up to 8192 tokens. The other models can process up to 512. Note that longer text inputs both make the mode slower and decrease performance, so if you're only working with texts of up to 400~ words / 1 page, use e.g. a deberta model for better performance. - The latest updates on new models are always available in the [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). ## Reproduction Reproduction code is available in the `v2_synthetic_data` directory here: https://github.com/MoritzLaurer/zeroshot-classifier/tree/main ## Limitations and bias The model can only do text classification tasks. Biases can come from the underlying foundation model, the human NLI training data and the synthetic data generated by Mixtral. ## License The foundation model was published under the MIT license. The licenses of the training data vary depending on the model, see above. ## Citation This model is an extension of the research described in this [paper](https://arxiv.org/pdf/2312.17543.pdf). If you use this model academically, please cite: ``` @misc{laurer_building_2023, title = {Building {Efficient} {Universal} {Classifiers} with {Natural} {Language} {Inference}}, url = {http://arxiv.org/abs/2312.17543}, doi = {10.48550/arXiv.2312.17543}, abstract = {Generative Large Language Models (LLMs) have become the mainstream choice for fewshot and zeroshot learning thanks to the universality of text generation. Many users, however, do not need the broad capabilities of generative LLMs when they only want to automate a classification task. Smaller BERT-like models can also learn universal tasks, which allow them to do any text classification task without requiring fine-tuning (zeroshot classification) or to learn new tasks with only a few examples (fewshot), while being significantly more efficient than generative LLMs. This paper (1) explains how Natural Language Inference (NLI) can be used as a universal classification task that follows similar principles as instruction fine-tuning of generative LLMs, (2) provides a step-by-step guide with reusable Jupyter notebooks for building a universal classifier, and (3) shares the resulting universal classifier that is trained on 33 datasets with 389 diverse classes. Parts of the code we share has been used to train our older zeroshot classifiers that have been downloaded more than 55 million times via the Hugging Face Hub as of December 2023. Our new classifier improves zeroshot performance by 9.4\%.}, urldate = {2024-01-05}, publisher = {arXiv}, author = {Laurer, Moritz and van Atteveldt, Wouter and Casas, Andreu and Welbers, Kasper}, month = dec, year = {2023}, note = {arXiv:2312.17543 [cs]}, keywords = {Computer Science - Artificial Intelligence, Computer Science - Computation and Language}, } ``` ### Ideas for cooperation or questions? If you have questions or ideas for cooperation, contact me at moritz{at}huggingface{dot}co or [LinkedIn](https://www.linkedin.com/in/moritz-laurer/) ### Flexible usage and "prompting" You can formulate your own hypotheses by changing the `hypothesis_template` of the zeroshot pipeline. Similar to "prompt engineering" for LLMs, you can test different formulations of your `hypothesis_template` and verbalized classes to improve performance. ```python from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" # formulation 1 hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] # formulation 2 depending on your use-case hypothesis_template = "The topic of this text is {}" classes_verbalized = ["political activities", "economic policy", "entertainment or music", "environmental protection"] # test different formulations zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ```	{"base_model": "microsoft/deberta-v3-base", "language": ["en"], "library_name": "transformers", "license": "mit", "pipeline_tag": "zero-shot-classification", "tags": ["text-classification", "zero-shot-classification"]}	task	[ "TEXT_CLASSIFICATION", "TRANSLATION" ]	44,534
cstr/whisper-large-v3-turbo-int8_float32	cstr	automatic-speech-recognition	[ "audio", "automatic-speech-recognition", "hf-asr-leaderboard", "en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", "no", "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su", "arxiv:2212.04356", "license:apache-2.0", "region:us" ]	2024-10-02T06:17:42Z	2024-10-02T07:23:38+00:00	57	0	--- language: - en - zh - de - es - ru - ko - fr - ja - pt - tr - pl - ca - nl - ar - sv - it - id - hi - fi - vi - he - uk - el - ms - cs - ro - da - hu - ta - false - th - ur - hr - bg - lt - la - mi - ml - cy - sk - te - fa - lv - bn - sr - az - sl - kn - et - mk - br - eu - is - hy - ne - mn - bs - kk - sq - sw - gl - mr - pa - si - km - sn - yo - so - af - oc - ka - be - tg - sd - gu - am - yi - lo - uz - fo - ht - ps - tk - nn - mt - sa - lb - my - bo - tl - mg - as - tt - haw - ln - ha - ba - jw - su license: apache-2.0 pipeline_tag: automatic-speech-recognition tags: - audio - automatic-speech-recognition - hf-asr-leaderboard widget: - example_title: Librispeech sample 1 src: https://cdn-media.huggingface.co/speech_samples/sample1.flac - example_title: Librispeech sample 2 src: https://cdn-media.huggingface.co/speech_samples/sample2.flac --- # Whisper turbo v3 for CTranslate2/faster-whisper This is an int8 quantified version of the OpenAI Whisper v3 turbo model. You can use it like this (example for usage in kaggle/Colab): ```kaggle # Clone the repository !git clone https://github.com/SYSTRAN/faster-whisper.git %cd faster-whisper # Install requirements !pip install -r requirements.txt # Import necessary modules from faster_whisper import WhisperModel from faster_whisper.transcribe import BatchedInferencePipeline # Initialize the model model = WhisperModel("cstr/whisper-large-v3-turbo-int8_float32", device="auto", compute_type="int8") batched_model = BatchedInferencePipeline(model=model) # Change to home directory and download audio %cd ~ !wget -c https://mcdn.podbean.com/mf/web/dir5wty678b6g4vg/HoP_453_-_The_Price_is_Right_-_Law_and_Economics_in_the_Second_Scholastic5yxzh.mp3 -O audio.mp3 # Benchmark transcription time import time import osstart_time = time.time() segments, info = batched_model.transcribe("audio.mp3", batch_size=16) end_time = time.time() # Print transcription for segment in segments: print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text)) # Print transcription info and benchmark results print(f"\nLanguage: {info.language}, Probability: {info.language_probability:.2f}") print(f"Duration: {info.duration:.2f}s, Duration after VAD: {info.duration_after_vad:.2f}s") transcription_time = end_time - start_time print(f"\nTranscription time: {transcription_time:.2f} seconds") # Calculate real-time factor real_time_factor = info.duration / transcription_time print(f"Real-time factor: {real_time_factor:.2f}x") # Print audio file size audio_file_size = os.path.getsize("audio.mp3") / (1024 * 1024) # Size in MB print(f"Audio file size: {audio_file_size:.2f} MB") ``` # Whisper (original Model Card follows) Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://huggingface.co/papers/2212.04356) by Alec Radford et al. from OpenAI. Trained on >5M hours of labeled data, Whisper demonstrates a strong ability to generalise to many datasets and domains in a zero-shot setting. Whisper large-v3-turbo is a distilled version of [Whisper large-v3](https://huggingface.co/openai/whisper-large-v3). In other words, it's the exact same model, except that the number of decoding layers have reduced from 32 to 4. As a result, the model is way faster, at the expense of a minor quality degradation. Disclaimer: Content for this model card has partly been written by the 🤗 Hugging Face team, and partly copied and pasted from the original model card. ## Usage Whisper large-v3-turbo is supported in Hugging Face 🤗 Transformers. To run the model, first install the Transformers library. For this example, we'll also install 🤗 Datasets to load toy audio dataset from the Hugging Face Hub, and 🤗 Accelerate to reduce the model loading time: ```bash pip install --upgrade pip pip install --upgrade transformers datasets[audio] accelerate ``` The model can be used with the [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) class to transcribe audios of arbitrary length: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` To transcribe a local audio file, simply pass the path to your audio file when you call the pipeline: ```python result = pipe("audio.mp3") ``` Multiple audio files can be transcribed in parallel by specifying them as a list and setting the `batch_size` parameter: ```python result = pipe(["audio_1.mp3", "audio_2.mp3"], batch_size=2) ``` Transformers is compatible with all Whisper decoding strategies, such as temperature fallback and condition on previous tokens. The following example demonstrates how to enable these heuristics: ```python generate_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } result = pipe(sample, generate_kwargs=generate_kwargs) ``` Whisper predicts the language of the source audio automatically. If the source audio language is known a-priori, it can be passed as an argument to the pipeline: ```python result = pipe(sample, generate_kwargs={"language": "english"}) ``` By default, Whisper performs the task of speech transcription, where the source audio language is the same as the target text language. To perform speech translation, where the target text is in English, set the task to `"translate"`: ```python result = pipe(sample, generate_kwargs={"task": "translate"}) ``` Finally, the model can be made to predict timestamps. For sentence-level timestamps, pass the `return_timestamps` argument: ```python result = pipe(sample, return_timestamps=True) print(result["chunks"]) ``` And for word-level timestamps: ```python result = pipe(sample, return_timestamps="word") print(result["chunks"]) ``` The above arguments can be used in isolation or in combination. For example, to perform the task of speech transcription where the source audio is in French, and we want to return sentence-level timestamps, the following can be used: ```python result = pipe(sample, return_timestamps=True, generate_kwargs={"language": "french", "task": "translate"}) print(result["chunks"]) ``` <details> <summary> For more control over the generation parameters, use the model + processor API directly: </summary> ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor from datasets import Audio, load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") dataset = dataset.cast_column("audio", Audio(processor.feature_extractor.sampling_rate)) sample = dataset[0]["audio"] inputs = processor( sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt", truncation=False, padding="longest", return_attention_mask=True, ) inputs = inputs.to(device, dtype=torch_dtype) gen_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } pred_ids = model.generate(inputs, gen_kwargs) pred_text = processor.batch_decode(pred_ids, skip_special_tokens=True, decode_with_timestamps=False) print(pred_text) ``` </details> ## Additional Speed & Memory Improvements You can apply additional speed and memory improvements to Whisper to further reduce the inference speed and VRAM requirements. ### Chunked Long-Form Whisper has a receptive field of 30-seconds. To transcribe audios longer than this, one of two long-form algorithms are required: 1. Sequential: uses a "sliding window" for buffered inference, transcribing 30-second slices one after the other 2. Chunked: splits long audio files into shorter ones (with a small overlap between segments), transcribes each segment independently, and stitches the resulting transcriptions at the boundaries The sequential long-form algorithm should be used in either of the following scenarios: 1. Transcription accuracy is the most important factor, and speed is less of a consideration 2. You are transcribing batches of long audio files, in which case the latency of sequential is comparable to chunked, while being up to 0.5% WER more accurate Conversely, the chunked algorithm should be used when: 1. Transcription speed is the most important factor 2. You are transcribing a single long audio file By default, Transformers uses the sequential algorithm. To enable the chunked algorithm, pass the `chunk_length_s` parameter to the `pipeline`. For large-v3, a chunk length of 30-seconds is optimal. To activate batching over long audio files, pass the argument `batch_size`: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, chunk_length_s=30, batch_size=16, # batch size for inference - set based on your device torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` #### Torch compile The Whisper forward pass is compatible with [`torch.compile`](https://pytorch.org/docs/stable/generated/torch.compile.html) for 4.5x speed-ups. Note: `torch.compile` is currently not compatible with the Chunked long-form algorithm or Flash Attention 2 ⚠️ ```python import torch from torch.nn.attention import SDPBackend, sdpa_kernel from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset from tqdm import tqdm torch.set_float32_matmul_precision("high") device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ).to(device) # Enable static cache and compile the forward pass model.generation_config.cache_implementation = "static" model.generation_config.max_new_tokens = 256 model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] # 2 warmup steps for _ in tqdm(range(2), desc="Warm-up step"): with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy(), generate_kwargs={"min_new_tokens": 256, "max_new_tokens": 256}) # fast run with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy()) print(result["text"]) ``` #### Flash Attention 2 We recommend using [Flash-Attention 2](https://huggingface.co/docs/transformers/main/en/perf_infer_gpu_one#flashattention-2) if your GPU supports it and you are not using [torch.compile](#torch-compile). To do so, first install [Flash Attention](https://github.com/Dao-AILab/flash-attention): ``` pip install flash-attn --no-build-isolation ``` Then pass `attn_implementation="flash_attention_2"` to `from_pretrained`: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="flash_attention_2") ``` #### Torch Scale-Product-Attention (SDPA) If your GPU does not support Flash Attention, we recommend making use of PyTorch [scaled dot-product attention (SDPA)](https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html). This attention implementation is activated by default for PyTorch versions 2.1.1 or greater. To check whether you have a compatible PyTorch version, run the following Python code snippet: ```python from transformers.utils import is_torch_sdpa_available print(is_torch_sdpa_available()) ``` If the above returns `True`, you have a valid version of PyTorch installed and SDPA is activated by default. If it returns `False`, you need to upgrade your PyTorch version according to the [official instructions](https://pytorch.org/get-started/locally/) Once a valid PyTorch version is installed, SDPA is activated by default. It can also be set explicitly by specifying `attn_implementation="sdpa"` as follows: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="sdpa") ``` For more information about how to use the SDPA refer to the [Transformers SDPA documentation](https://huggingface.co/docs/transformers/en/perf_infer_gpu_one#pytorch-scaled-dot-product-attention). ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. There are two flavours of Whisper model: English-only and multilingual. The English-only models were trained on the task of English speech recognition. The multilingual models were trained simultaneously on multilingual speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the same language as the audio. For speech translation, the model predicts transcriptions to a different language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are available as English-only and multilingual. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: \| Size \| Parameters \| English-only \| Multilingual \| \|----------\|------------\|------------------------------------------------------\|-----------------------------------------------------\| \| tiny \| 39 M \| [✓](https://huggingface.co/openai/whisper-tiny.en) \| [✓](https://huggingface.co/openai/whisper-tiny) \| \| base \| 74 M \| [✓](https://huggingface.co/openai/whisper-base.en) \| [✓](https://huggingface.co/openai/whisper-base) \| \| small \| 244 M \| [✓](https://huggingface.co/openai/whisper-small.en) \| [✓](https://huggingface.co/openai/whisper-small) \| \| medium \| 769 M \| [✓](https://huggingface.co/openai/whisper-medium.en) \| [✓](https://huggingface.co/openai/whisper-medium) \| \| large \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large) \| \| large-v2 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v2) \| \| large-v3 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v3) \| \| large-v3-turbo \| 809 M \| x \| [✓](https://huggingface.co/ylacombe/whisper-large-v3-turbo) \| ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through fine-tuning. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data No information provided. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```	null	Non_BioNLP	# Whisper turbo v3 for CTranslate2/faster-whisper This is an int8 quantified version of the OpenAI Whisper v3 turbo model. You can use it like this (example for usage in kaggle/Colab): ```kaggle # Clone the repository !git clone https://github.com/SYSTRAN/faster-whisper.git %cd faster-whisper # Install requirements !pip install -r requirements.txt # Import necessary modules from faster_whisper import WhisperModel from faster_whisper.transcribe import BatchedInferencePipeline # Initialize the model model = WhisperModel("cstr/whisper-large-v3-turbo-int8_float32", device="auto", compute_type="int8") batched_model = BatchedInferencePipeline(model=model) # Change to home directory and download audio %cd ~ !wget -c https://mcdn.podbean.com/mf/web/dir5wty678b6g4vg/HoP_453_-_The_Price_is_Right_-_Law_and_Economics_in_the_Second_Scholastic5yxzh.mp3 -O audio.mp3 # Benchmark transcription time import time import osstart_time = time.time() segments, info = batched_model.transcribe("audio.mp3", batch_size=16) end_time = time.time() # Print transcription for segment in segments: print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text)) # Print transcription info and benchmark results print(f"\nLanguage: {info.language}, Probability: {info.language_probability:.2f}") print(f"Duration: {info.duration:.2f}s, Duration after VAD: {info.duration_after_vad:.2f}s") transcription_time = end_time - start_time print(f"\nTranscription time: {transcription_time:.2f} seconds") # Calculate real-time factor real_time_factor = info.duration / transcription_time print(f"Real-time factor: {real_time_factor:.2f}x") # Print audio file size audio_file_size = os.path.getsize("audio.mp3") / (1024 * 1024) # Size in MB print(f"Audio file size: {audio_file_size:.2f} MB") ``` # Whisper (original Model Card follows) Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://huggingface.co/papers/2212.04356) by Alec Radford et al. from OpenAI. Trained on >5M hours of labeled data, Whisper demonstrates a strong ability to generalise to many datasets and domains in a zero-shot setting. Whisper large-v3-turbo is a distilled version of [Whisper large-v3](https://huggingface.co/openai/whisper-large-v3). In other words, it's the exact same model, except that the number of decoding layers have reduced from 32 to 4. As a result, the model is way faster, at the expense of a minor quality degradation. Disclaimer: Content for this model card has partly been written by the 🤗 Hugging Face team, and partly copied and pasted from the original model card. ## Usage Whisper large-v3-turbo is supported in Hugging Face 🤗 Transformers. To run the model, first install the Transformers library. For this example, we'll also install 🤗 Datasets to load toy audio dataset from the Hugging Face Hub, and 🤗 Accelerate to reduce the model loading time: ```bash pip install --upgrade pip pip install --upgrade transformers datasets[audio] accelerate ``` The model can be used with the [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) class to transcribe audios of arbitrary length: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` To transcribe a local audio file, simply pass the path to your audio file when you call the pipeline: ```python result = pipe("audio.mp3") ``` Multiple audio files can be transcribed in parallel by specifying them as a list and setting the `batch_size` parameter: ```python result = pipe(["audio_1.mp3", "audio_2.mp3"], batch_size=2) ``` Transformers is compatible with all Whisper decoding strategies, such as temperature fallback and condition on previous tokens. The following example demonstrates how to enable these heuristics: ```python generate_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } result = pipe(sample, generate_kwargs=generate_kwargs) ``` Whisper predicts the language of the source audio automatically. If the source audio language is known a-priori, it can be passed as an argument to the pipeline: ```python result = pipe(sample, generate_kwargs={"language": "english"}) ``` By default, Whisper performs the task of speech transcription, where the source audio language is the same as the target text language. To perform speech translation, where the target text is in English, set the task to `"translate"`: ```python result = pipe(sample, generate_kwargs={"task": "translate"}) ``` Finally, the model can be made to predict timestamps. For sentence-level timestamps, pass the `return_timestamps` argument: ```python result = pipe(sample, return_timestamps=True) print(result["chunks"]) ``` And for word-level timestamps: ```python result = pipe(sample, return_timestamps="word") print(result["chunks"]) ``` The above arguments can be used in isolation or in combination. For example, to perform the task of speech transcription where the source audio is in French, and we want to return sentence-level timestamps, the following can be used: ```python result = pipe(sample, return_timestamps=True, generate_kwargs={"language": "french", "task": "translate"}) print(result["chunks"]) ``` <details> <summary> For more control over the generation parameters, use the model + processor API directly: </summary> ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor from datasets import Audio, load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") dataset = dataset.cast_column("audio", Audio(processor.feature_extractor.sampling_rate)) sample = dataset[0]["audio"] inputs = processor( sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt", truncation=False, padding="longest", return_attention_mask=True, ) inputs = inputs.to(device, dtype=torch_dtype) gen_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } pred_ids = model.generate(inputs, gen_kwargs) pred_text = processor.batch_decode(pred_ids, skip_special_tokens=True, decode_with_timestamps=False) print(pred_text) ``` </details> ## Additional Speed & Memory Improvements You can apply additional speed and memory improvements to Whisper to further reduce the inference speed and VRAM requirements. ### Chunked Long-Form Whisper has a receptive field of 30-seconds. To transcribe audios longer than this, one of two long-form algorithms are required: 1. Sequential: uses a "sliding window" for buffered inference, transcribing 30-second slices one after the other 2. Chunked: splits long audio files into shorter ones (with a small overlap between segments), transcribes each segment independently, and stitches the resulting transcriptions at the boundaries The sequential long-form algorithm should be used in either of the following scenarios: 1. Transcription accuracy is the most important factor, and speed is less of a consideration 2. You are transcribing batches of long audio files, in which case the latency of sequential is comparable to chunked, while being up to 0.5% WER more accurate Conversely, the chunked algorithm should be used when: 1. Transcription speed is the most important factor 2. You are transcribing a single long audio file By default, Transformers uses the sequential algorithm. To enable the chunked algorithm, pass the `chunk_length_s` parameter to the `pipeline`. For large-v3, a chunk length of 30-seconds is optimal. To activate batching over long audio files, pass the argument `batch_size`: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, chunk_length_s=30, batch_size=16, # batch size for inference - set based on your device torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` #### Torch compile The Whisper forward pass is compatible with [`torch.compile`](https://pytorch.org/docs/stable/generated/torch.compile.html) for 4.5x speed-ups. Note: `torch.compile` is currently not compatible with the Chunked long-form algorithm or Flash Attention 2 ⚠️ ```python import torch from torch.nn.attention import SDPBackend, sdpa_kernel from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset from tqdm import tqdm torch.set_float32_matmul_precision("high") device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ).to(device) # Enable static cache and compile the forward pass model.generation_config.cache_implementation = "static" model.generation_config.max_new_tokens = 256 model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] # 2 warmup steps for _ in tqdm(range(2), desc="Warm-up step"): with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy(), generate_kwargs={"min_new_tokens": 256, "max_new_tokens": 256}) # fast run with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy()) print(result["text"]) ``` #### Flash Attention 2 We recommend using [Flash-Attention 2](https://huggingface.co/docs/transformers/main/en/perf_infer_gpu_one#flashattention-2) if your GPU supports it and you are not using [torch.compile](#torch-compile). To do so, first install [Flash Attention](https://github.com/Dao-AILab/flash-attention): ``` pip install flash-attn --no-build-isolation ``` Then pass `attn_implementation="flash_attention_2"` to `from_pretrained`: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="flash_attention_2") ``` #### Torch Scale-Product-Attention (SDPA) If your GPU does not support Flash Attention, we recommend making use of PyTorch [scaled dot-product attention (SDPA)](https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html). This attention implementation is activated by default for PyTorch versions 2.1.1 or greater. To check whether you have a compatible PyTorch version, run the following Python code snippet: ```python from transformers.utils import is_torch_sdpa_available print(is_torch_sdpa_available()) ``` If the above returns `True`, you have a valid version of PyTorch installed and SDPA is activated by default. If it returns `False`, you need to upgrade your PyTorch version according to the [official instructions](https://pytorch.org/get-started/locally/) Once a valid PyTorch version is installed, SDPA is activated by default. It can also be set explicitly by specifying `attn_implementation="sdpa"` as follows: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="sdpa") ``` For more information about how to use the SDPA refer to the [Transformers SDPA documentation](https://huggingface.co/docs/transformers/en/perf_infer_gpu_one#pytorch-scaled-dot-product-attention). ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. There are two flavours of Whisper model: English-only and multilingual. The English-only models were trained on the task of English speech recognition. The multilingual models were trained simultaneously on multilingual speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the same language as the audio. For speech translation, the model predicts transcriptions to a different language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are available as English-only and multilingual. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: \| Size \| Parameters \| English-only \| Multilingual \| \|----------\|------------\|------------------------------------------------------\|-----------------------------------------------------\| \| tiny \| 39 M \| [✓](https://huggingface.co/openai/whisper-tiny.en) \| [✓](https://huggingface.co/openai/whisper-tiny) \| \| base \| 74 M \| [✓](https://huggingface.co/openai/whisper-base.en) \| [✓](https://huggingface.co/openai/whisper-base) \| \| small \| 244 M \| [✓](https://huggingface.co/openai/whisper-small.en) \| [✓](https://huggingface.co/openai/whisper-small) \| \| medium \| 769 M \| [✓](https://huggingface.co/openai/whisper-medium.en) \| [✓](https://huggingface.co/openai/whisper-medium) \| \| large \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large) \| \| large-v2 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v2) \| \| large-v3 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v3) \| \| large-v3-turbo \| 809 M \| x \| [✓](https://huggingface.co/ylacombe/whisper-large-v3-turbo) \| ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through fine-tuning. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data No information provided. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```	{"language": ["en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", false, "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su"], "license": "apache-2.0", "pipeline_tag": "automatic-speech-recognition", "tags": ["audio", "automatic-speech-recognition", "hf-asr-leaderboard"], "widget": [{"example_title": "Librispeech sample 1", "src": "https://cdn-media.huggingface.co/speech_samples/sample1.flac"}, {"example_title": "Librispeech sample 2", "src": "https://cdn-media.huggingface.co/speech_samples/sample2.flac"}]}	task	[ "TRANSLATION" ]	44,535
Helsinki-NLP/opus-mt-tc-big-gmq-itc	Helsinki-NLP	translation	[ "transformers", "pytorch", "tf", "safetensors", "marian", "text2text-generation", "translation", "opus-mt-tc", "ca", "da", "es", "fr", "gl", "is", "it", "nb", "pt", "ro", "sv", "license:cc-by-4.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-08-12T11:15:27Z	2023-10-10T11:14:55+00:00	24	0	--- language: - ca - da - es - fr - gl - is - it - nb - pt - ro - sv license: cc-by-4.0 tags: - translation - opus-mt-tc model-index: - name: opus-mt-tc-big-gmq-itc results: - task: type: translation name: Translation dan-cat dataset: name: flores101-devtest type: flores_101 args: dan cat devtest metrics: - type: bleu value: 33.4 name: BLEU - type: chrf value: 0.59224 name: chr-F - type: bleu value: 38.3 name: BLEU - type: chrf value: 0.63387 name: chr-F - type: bleu value: 26.4 name: BLEU - type: chrf value: 0.54446 name: chr-F - type: bleu value: 25.7 name: BLEU - type: chrf value: 0.55237 name: chr-F - type: bleu value: 36.9 name: BLEU - type: chrf value: 0.62233 name: chr-F - type: bleu value: 31.8 name: BLEU - type: chrf value: 0.58235 name: chr-F - type: bleu value: 24.3 name: BLEU - type: chrf value: 0.52453 name: chr-F - type: bleu value: 22.7 name: BLEU - type: chrf value: 0.4893 name: chr-F - type: bleu value: 26.2 name: BLEU - type: chrf value: 0.52704 name: chr-F - type: bleu value: 18.0 name: BLEU - type: chrf value: 0.45387 name: chr-F - type: bleu value: 18.6 name: BLEU - type: chrf value: 0.47303 name: chr-F - type: bleu value: 24.9 name: BLEU - type: chrf value: 0.51381 name: chr-F - type: bleu value: 21.6 name: BLEU - type: chrf value: 0.48224 name: chr-F - type: bleu value: 18.1 name: BLEU - type: chrf value: 0.45786 name: chr-F - type: bleu value: 28.9 name: BLEU - type: chrf value: 0.55984 name: chr-F - type: bleu value: 33.8 name: BLEU - type: chrf value: 0.60102 name: chr-F - type: bleu value: 23.4 name: BLEU - type: chrf value: 0.52145 name: chr-F - type: bleu value: 22.2 name: BLEU - type: chrf value: 0.52619 name: chr-F - type: bleu value: 32.2 name: BLEU - type: chrf value: 0.58836 name: chr-F - type: bleu value: 27.6 name: BLEU - type: chrf value: 0.54845 name: chr-F - type: bleu value: 21.8 name: BLEU - type: chrf value: 0.50661 name: chr-F - type: bleu value: 32.4 name: BLEU - type: chrf value: 0.58542 name: chr-F - type: bleu value: 39.3 name: BLEU - type: chrf value: 0.63688 name: chr-F - type: bleu value: 26.0 name: BLEU - type: chrf value: 0.53989 name: chr-F - type: bleu value: 25.9 name: BLEU - type: chrf value: 0.55232 name: chr-F - type: bleu value: 36.5 name: BLEU - type: chrf value: 0.61882 name: chr-F - type: bleu value: 31.0 name: BLEU - type: chrf value: 0.57419 name: chr-F - type: bleu value: 23.8 name: BLEU - type: chrf value: 0.52175 name: chr-F - task: type: translation name: Translation dan-fra dataset: name: tatoeba-test-v2021-08-07 type: tatoeba_mt args: dan-fra metrics: - type: bleu value: 63.8 name: BLEU - type: chrf value: 0.76671 name: chr-F - type: bleu value: 56.2 name: BLEU - type: chrf value: 0.74658 name: chr-F - type: bleu value: 57.8 name: BLEU - type: chrf value: 0.74944 name: chr-F - type: bleu value: 54.8 name: BLEU - type: chrf value: 0.72328 name: chr-F - type: bleu value: 51.0 name: BLEU - type: chrf value: 0.69354 name: chr-F - type: bleu value: 49.2 name: BLEU - type: chrf value: 0.66008 name: chr-F - type: bleu value: 54.4 name: BLEU - type: chrf value: 0.70854 name: chr-F - type: bleu value: 55.9 name: BLEU - type: chrf value: 0.73672 name: chr-F - type: bleu value: 59.2 name: BLEU - type: chrf value: 0.73014 name: chr-F - type: bleu value: 56.6 name: BLEU - type: chrf value: 0.73211 name: chr-F - type: bleu value: 48.7 name: BLEU - type: chrf value: 0.68146 name: chr-F - type: bleu value: 55.3 name: BLEU - type: chrf value: 0.71373 name: chr-F --- # opus-mt-tc-big-gmq-itc ## Table of Contents - [Model Details](#model-details) - [Uses](#uses) - [Risks, Limitations and Biases](#risks-limitations-and-biases) - [How to Get Started With the Model](#how-to-get-started-with-the-model) - [Training](#training) - [Evaluation](#evaluation) - [Citation Information](#citation-information) - [Acknowledgements](#acknowledgements) ## Model Details Neural machine translation model for translating from North Germanic languages (gmq) to Italic languages (itc). This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train). Model Description: - Developed by: Language Technology Research Group at the University of Helsinki - Model Type: Translation (transformer-big) - Release: 2022-08-09 - License: CC-BY-4.0 - Language(s): - Source Language(s): dan isl nno nob nor swe - Target Language(s): cat fra glg ita lat por ron spa - Language Pair(s): dan-cat dan-fra dan-glg dan-ita dan-por dan-ron dan-spa isl-cat isl-fra isl-ita isl-por isl-ron isl-spa nob-cat nob-fra nob-glg nob-ita nob-por nob-ron nob-spa swe-cat swe-fra swe-glg swe-ita swe-por swe-ron swe-spa - Valid Target Language Labels: >>acf<< >>aoa<< >>arg<< >>ast<< >>cat<< >>cbk<< >>ccd<< >>cks<< >>cos<< >>cri<< >>crs<< >>dlm<< >>drc<< >>egl<< >>ext<< >>fab<< >>fax<< >>fra<< >>frc<< >>frm<< >>fro<< >>frp<< >>fur<< >>gcf<< >>gcr<< >>glg<< >>hat<< >>idb<< >>ist<< >>ita<< >>itk<< >>kea<< >>kmv<< >>lad<< >>lad_Latn<< >>lat<< >>lat_Latn<< >>lij<< >>lld<< >>lmo<< >>lou<< >>mcm<< >>mfe<< >>mol<< >>mwl<< >>mxi<< >>mzs<< >>nap<< >>nrf<< >>oci<< >>osc<< >>osp<< >>osp_Latn<< >>pap<< >>pcd<< >>pln<< >>pms<< >>pob<< >>por<< >>pov<< >>pre<< >>pro<< >>qbb<< >>qhr<< >>rcf<< >>rgn<< >>roh<< >>ron<< >>ruo<< >>rup<< >>ruq<< >>scf<< >>scn<< >>sdc<< >>sdn<< >>spa<< >>spq<< >>spx<< >>src<< >>srd<< >>sro<< >>tmg<< >>tvy<< >>vec<< >>vkp<< >>wln<< >>xfa<< >>xum<< - Original Model: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - Resources for more information: - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) - More information about released models for this language pair: [OPUS-MT gmq-itc README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/gmq-itc/README.md) - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian) - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/ This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>fra<<` ## Uses This model can be used for translation and text-to-text generation. ## Risks, Limitations and Biases CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes. Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). ## How to Get Started With the Model A short example code: ```python from transformers import MarianMTModel, MarianTokenizer src_text = [ ">>spa<< Jag är inte religiös.", ">>por<< Livet er for kort til å lære seg tysk." ] model_name = "pytorch-models/opus-mt-tc-big-gmq-itc" tokenizer = MarianTokenizer.from_pretrained(model_name) model = MarianMTModel.from_pretrained(model_name) translated = model.generate(tokenizer(src_text, return_tensors="pt", padding=True)) for t in translated: print( tokenizer.decode(t, skip_special_tokens=True) ) # expected output: # No soy religioso. # A vida é muito curta para aprender alemão. ``` You can also use OPUS-MT models with the transformers pipelines, for example: ```python from transformers import pipeline pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-gmq-itc") print(pipe(">>spa<< Jag är inte religiös.")) # expected output: No soy religioso. ``` ## Training - Data: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge)) - Pre-processing: SentencePiece (spm32k,spm32k) - Model Type: transformer-big - Original MarianNMT Model: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - Training Scripts*: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) ## Evaluation test set translations: [opusTCv20210807_transformer-big_2022-08-09.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.test.txt) * test set scores: [opusTCv20210807_transformer-big_2022-08-09.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.eval.txt) * benchmark results: [benchmark_results.txt](benchmark_results.txt) * benchmark output: [benchmark_translations.zip](benchmark_translations.zip) \| langpair \| testset \| chr-F \| BLEU \| #sent \| #words \| \|----------\|---------\|-------\|-------\|-------\|--------\| \| dan-fra \| tatoeba-test-v2021-08-07 \| 0.76671 \| 63.8 \| 1731 \| 11882 \| \| dan-ita \| tatoeba-test-v2021-08-07 \| 0.74658 \| 56.2 \| 284 \| 2226 \| \| dan-por \| tatoeba-test-v2021-08-07 \| 0.74944 \| 57.8 \| 873 \| 5360 \| \| dan-spa \| tatoeba-test-v2021-08-07 \| 0.72328 \| 54.8 \| 5000 \| 35528 \| \| isl-ita \| tatoeba-test-v2021-08-07 \| 0.69354 \| 51.0 \| 236 \| 1450 \| \| isl-spa \| tatoeba-test-v2021-08-07 \| 0.66008 \| 49.2 \| 238 \| 1229 \| \| nob-fra \| tatoeba-test-v2021-08-07 \| 0.70854 \| 54.4 \| 323 \| 2269 \| \| nob-spa \| tatoeba-test-v2021-08-07 \| 0.73672 \| 55.9 \| 885 \| 6866 \| \| swe-fra \| tatoeba-test-v2021-08-07 \| 0.73014 \| 59.2 \| 1407 \| 9580 \| \| swe-ita \| tatoeba-test-v2021-08-07 \| 0.73211 \| 56.6 \| 715 \| 4711 \| \| swe-por \| tatoeba-test-v2021-08-07 \| 0.68146 \| 48.7 \| 320 \| 2032 \| \| swe-spa \| tatoeba-test-v2021-08-07 \| 0.71373 \| 55.3 \| 1351 \| 8235 \| \| dan-cat \| flores101-devtest \| 0.59224 \| 33.4 \| 1012 \| 27304 \| \| dan-fra \| flores101-devtest \| 0.63387 \| 38.3 \| 1012 \| 28343 \| \| dan-glg \| flores101-devtest \| 0.54446 \| 26.4 \| 1012 \| 26582 \| \| dan-ita \| flores101-devtest \| 0.55237 \| 25.7 \| 1012 \| 27306 \| \| dan-por \| flores101-devtest \| 0.62233 \| 36.9 \| 1012 \| 26519 \| \| dan-ron \| flores101-devtest \| 0.58235 \| 31.8 \| 1012 \| 26799 \| \| dan-spa \| flores101-devtest \| 0.52453 \| 24.3 \| 1012 \| 29199 \| \| isl-cat \| flores101-devtest \| 0.48930 \| 22.7 \| 1012 \| 27304 \| \| isl-fra \| flores101-devtest \| 0.52704 \| 26.2 \| 1012 \| 28343 \| \| isl-glg \| flores101-devtest \| 0.45387 \| 18.0 \| 1012 \| 26582 \| \| isl-ita \| flores101-devtest \| 0.47303 \| 18.6 \| 1012 \| 27306 \| \| isl-por \| flores101-devtest \| 0.51381 \| 24.9 \| 1012 \| 26519 \| \| isl-ron \| flores101-devtest \| 0.48224 \| 21.6 \| 1012 \| 26799 \| \| isl-spa \| flores101-devtest \| 0.45786 \| 18.1 \| 1012 \| 29199 \| \| nob-cat \| flores101-devtest \| 0.55984 \| 28.9 \| 1012 \| 27304 \| \| nob-fra \| flores101-devtest \| 0.60102 \| 33.8 \| 1012 \| 28343 \| \| nob-glg \| flores101-devtest \| 0.52145 \| 23.4 \| 1012 \| 26582 \| \| nob-ita \| flores101-devtest \| 0.52619 \| 22.2 \| 1012 \| 27306 \| \| nob-por \| flores101-devtest \| 0.58836 \| 32.2 \| 1012 \| 26519 \| \| nob-ron \| flores101-devtest \| 0.54845 \| 27.6 \| 1012 \| 26799 \| \| nob-spa \| flores101-devtest \| 0.50661 \| 21.8 \| 1012 \| 29199 \| \| swe-cat \| flores101-devtest \| 0.58542 \| 32.4 \| 1012 \| 27304 \| \| swe-fra \| flores101-devtest \| 0.63688 \| 39.3 \| 1012 \| 28343 \| \| swe-glg \| flores101-devtest \| 0.53989 \| 26.0 \| 1012 \| 26582 \| \| swe-ita \| flores101-devtest \| 0.55232 \| 25.9 \| 1012 \| 27306 \| \| swe-por \| flores101-devtest \| 0.61882 \| 36.5 \| 1012 \| 26519 \| \| swe-ron \| flores101-devtest \| 0.57419 \| 31.0 \| 1012 \| 26799 \| \| swe-spa \| flores101-devtest \| 0.52175 \| 23.8 \| 1012 \| 29199 \| ## Citation Information * Publications: [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.) ``` @inproceedings{tiedemann-thottingal-2020-opus, title = "{OPUS}-{MT} {--} Building open translation services for the World", author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh}, booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation", month = nov, year = "2020", address = "Lisboa, Portugal", publisher = "European Association for Machine Translation", url = "https://aclanthology.org/2020.eamt-1.61", pages = "479--480", } @inproceedings{tiedemann-2020-tatoeba, title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}", author = {Tiedemann, J{\"o}rg}, booktitle = "Proceedings of the Fifth Conference on Machine Translation", month = nov, year = "2020", address = "Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2020.wmt-1.139", pages = "1174--1182", } ``` ## Acknowledgements The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland. ## Model conversion info * transformers version: 4.16.2 * OPUS-MT git hash: 8b9f0b0 * port time: Sat Aug 13 00:00:00 EEST 2022 * port machine: LM0-400-22516.local	null	Non_BioNLP	# opus-mt-tc-big-gmq-itc ## Table of Contents - [Model Details](#model-details) - [Uses](#uses) - [Risks, Limitations and Biases](#risks-limitations-and-biases) - [How to Get Started With the Model](#how-to-get-started-with-the-model) - [Training](#training) - [Evaluation](#evaluation) - [Citation Information](#citation-information) - [Acknowledgements](#acknowledgements) ## Model Details Neural machine translation model for translating from North Germanic languages (gmq) to Italic languages (itc). This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train). Model Description: - Developed by: Language Technology Research Group at the University of Helsinki - Model Type: Translation (transformer-big) - Release: 2022-08-09 - License: CC-BY-4.0 - Language(s): - Source Language(s): dan isl nno nob nor swe - Target Language(s): cat fra glg ita lat por ron spa - Language Pair(s): dan-cat dan-fra dan-glg dan-ita dan-por dan-ron dan-spa isl-cat isl-fra isl-ita isl-por isl-ron isl-spa nob-cat nob-fra nob-glg nob-ita nob-por nob-ron nob-spa swe-cat swe-fra swe-glg swe-ita swe-por swe-ron swe-spa - Valid Target Language Labels: >>acf<< >>aoa<< >>arg<< >>ast<< >>cat<< >>cbk<< >>ccd<< >>cks<< >>cos<< >>cri<< >>crs<< >>dlm<< >>drc<< >>egl<< >>ext<< >>fab<< >>fax<< >>fra<< >>frc<< >>frm<< >>fro<< >>frp<< >>fur<< >>gcf<< >>gcr<< >>glg<< >>hat<< >>idb<< >>ist<< >>ita<< >>itk<< >>kea<< >>kmv<< >>lad<< >>lad_Latn<< >>lat<< >>lat_Latn<< >>lij<< >>lld<< >>lmo<< >>lou<< >>mcm<< >>mfe<< >>mol<< >>mwl<< >>mxi<< >>mzs<< >>nap<< >>nrf<< >>oci<< >>osc<< >>osp<< >>osp_Latn<< >>pap<< >>pcd<< >>pln<< >>pms<< >>pob<< >>por<< >>pov<< >>pre<< >>pro<< >>qbb<< >>qhr<< >>rcf<< >>rgn<< >>roh<< >>ron<< >>ruo<< >>rup<< >>ruq<< >>scf<< >>scn<< >>sdc<< >>sdn<< >>spa<< >>spq<< >>spx<< >>src<< >>srd<< >>sro<< >>tmg<< >>tvy<< >>vec<< >>vkp<< >>wln<< >>xfa<< >>xum<< - Original Model: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - Resources for more information: - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) - More information about released models for this language pair: [OPUS-MT gmq-itc README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/gmq-itc/README.md) - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian) - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/ This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>fra<<` ## Uses This model can be used for translation and text-to-text generation. ## Risks, Limitations and Biases CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes. Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). ## How to Get Started With the Model A short example code: ```python from transformers import MarianMTModel, MarianTokenizer src_text = [ ">>spa<< Jag är inte religiös.", ">>por<< Livet er for kort til å lære seg tysk." ] model_name = "pytorch-models/opus-mt-tc-big-gmq-itc" tokenizer = MarianTokenizer.from_pretrained(model_name) model = MarianMTModel.from_pretrained(model_name) translated = model.generate(tokenizer(src_text, return_tensors="pt", padding=True)) for t in translated: print( tokenizer.decode(t, skip_special_tokens=True) ) # expected output: # No soy religioso. # A vida é muito curta para aprender alemão. ``` You can also use OPUS-MT models with the transformers pipelines, for example: ```python from transformers import pipeline pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-gmq-itc") print(pipe(">>spa<< Jag är inte religiös.")) # expected output: No soy religioso. ``` ## Training - Data: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge)) - Pre-processing: SentencePiece (spm32k,spm32k) - Model Type: transformer-big - Original MarianNMT Model: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - Training Scripts*: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) ## Evaluation test set translations: [opusTCv20210807_transformer-big_2022-08-09.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.test.txt) * test set scores: [opusTCv20210807_transformer-big_2022-08-09.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.eval.txt) * benchmark results: [benchmark_results.txt](benchmark_results.txt) * benchmark output: [benchmark_translations.zip](benchmark_translations.zip) \| langpair \| testset \| chr-F \| BLEU \| #sent \| #words \| \|----------\|---------\|-------\|-------\|-------\|--------\| \| dan-fra \| tatoeba-test-v2021-08-07 \| 0.76671 \| 63.8 \| 1731 \| 11882 \| \| dan-ita \| tatoeba-test-v2021-08-07 \| 0.74658 \| 56.2 \| 284 \| 2226 \| \| dan-por \| tatoeba-test-v2021-08-07 \| 0.74944 \| 57.8 \| 873 \| 5360 \| \| dan-spa \| tatoeba-test-v2021-08-07 \| 0.72328 \| 54.8 \| 5000 \| 35528 \| \| isl-ita \| tatoeba-test-v2021-08-07 \| 0.69354 \| 51.0 \| 236 \| 1450 \| \| isl-spa \| tatoeba-test-v2021-08-07 \| 0.66008 \| 49.2 \| 238 \| 1229 \| \| nob-fra \| tatoeba-test-v2021-08-07 \| 0.70854 \| 54.4 \| 323 \| 2269 \| \| nob-spa \| tatoeba-test-v2021-08-07 \| 0.73672 \| 55.9 \| 885 \| 6866 \| \| swe-fra \| tatoeba-test-v2021-08-07 \| 0.73014 \| 59.2 \| 1407 \| 9580 \| \| swe-ita \| tatoeba-test-v2021-08-07 \| 0.73211 \| 56.6 \| 715 \| 4711 \| \| swe-por \| tatoeba-test-v2021-08-07 \| 0.68146 \| 48.7 \| 320 \| 2032 \| \| swe-spa \| tatoeba-test-v2021-08-07 \| 0.71373 \| 55.3 \| 1351 \| 8235 \| \| dan-cat \| flores101-devtest \| 0.59224 \| 33.4 \| 1012 \| 27304 \| \| dan-fra \| flores101-devtest \| 0.63387 \| 38.3 \| 1012 \| 28343 \| \| dan-glg \| flores101-devtest \| 0.54446 \| 26.4 \| 1012 \| 26582 \| \| dan-ita \| flores101-devtest \| 0.55237 \| 25.7 \| 1012 \| 27306 \| \| dan-por \| flores101-devtest \| 0.62233 \| 36.9 \| 1012 \| 26519 \| \| dan-ron \| flores101-devtest \| 0.58235 \| 31.8 \| 1012 \| 26799 \| \| dan-spa \| flores101-devtest \| 0.52453 \| 24.3 \| 1012 \| 29199 \| \| isl-cat \| flores101-devtest \| 0.48930 \| 22.7 \| 1012 \| 27304 \| \| isl-fra \| flores101-devtest \| 0.52704 \| 26.2 \| 1012 \| 28343 \| \| isl-glg \| flores101-devtest \| 0.45387 \| 18.0 \| 1012 \| 26582 \| \| isl-ita \| flores101-devtest \| 0.47303 \| 18.6 \| 1012 \| 27306 \| \| isl-por \| flores101-devtest \| 0.51381 \| 24.9 \| 1012 \| 26519 \| \| isl-ron \| flores101-devtest \| 0.48224 \| 21.6 \| 1012 \| 26799 \| \| isl-spa \| flores101-devtest \| 0.45786 \| 18.1 \| 1012 \| 29199 \| \| nob-cat \| flores101-devtest \| 0.55984 \| 28.9 \| 1012 \| 27304 \| \| nob-fra \| flores101-devtest \| 0.60102 \| 33.8 \| 1012 \| 28343 \| \| nob-glg \| flores101-devtest \| 0.52145 \| 23.4 \| 1012 \| 26582 \| \| nob-ita \| flores101-devtest \| 0.52619 \| 22.2 \| 1012 \| 27306 \| \| nob-por \| flores101-devtest \| 0.58836 \| 32.2 \| 1012 \| 26519 \| \| nob-ron \| flores101-devtest \| 0.54845 \| 27.6 \| 1012 \| 26799 \| \| nob-spa \| flores101-devtest \| 0.50661 \| 21.8 \| 1012 \| 29199 \| \| swe-cat \| flores101-devtest \| 0.58542 \| 32.4 \| 1012 \| 27304 \| \| swe-fra \| flores101-devtest \| 0.63688 \| 39.3 \| 1012 \| 28343 \| \| swe-glg \| flores101-devtest \| 0.53989 \| 26.0 \| 1012 \| 26582 \| \| swe-ita \| flores101-devtest \| 0.55232 \| 25.9 \| 1012 \| 27306 \| \| swe-por \| flores101-devtest \| 0.61882 \| 36.5 \| 1012 \| 26519 \| \| swe-ron \| flores101-devtest \| 0.57419 \| 31.0 \| 1012 \| 26799 \| \| swe-spa \| flores101-devtest \| 0.52175 \| 23.8 \| 1012 \| 29199 \| ## Citation Information * Publications: [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.) ``` @inproceedings{tiedemann-thottingal-2020-opus, title = "{OPUS}-{MT} {--} Building open translation services for the World", author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh}, booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation", month = nov, year = "2020", address = "Lisboa, Portugal", publisher = "European Association for Machine Translation", url = "https://aclanthology.org/2020.eamt-1.61", pages = "479--480", } @inproceedings{tiedemann-2020-tatoeba, title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}", author = {Tiedemann, J{\"o}rg}, booktitle = "Proceedings of the Fifth Conference on Machine Translation", month = nov, year = "2020", address = "Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2020.wmt-1.139", pages = "1174--1182", } ``` ## Acknowledgements The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland. ## Model conversion info * transformers version: 4.16.2 * OPUS-MT git hash: 8b9f0b0 * port time: Sat Aug 13 00:00:00 EEST 2022 * port machine: LM0-400-22516.local	{"language": ["ca", "da", "es", "fr", "gl", "is", "it", "nb", "pt", "ro", "sv"], "license": "cc-by-4.0", "tags": ["translation", "opus-mt-tc"], "model-index": [{"name": "opus-mt-tc-big-gmq-itc", "results": [{"task": {"type": "translation", "name": "Translation dan-cat"}, "dataset": {"name": "flores101-devtest", "type": "flores_101", "args": "dan cat devtest"}, "metrics": [{"type": "bleu", "value": 33.4, "name": "BLEU"}, {"type": "chrf", "value": 0.59224, "name": "chr-F"}, {"type": "bleu", "value": 38.3, "name": "BLEU"}, {"type": "chrf", "value": 0.63387, "name": "chr-F"}, {"type": "bleu", "value": 26.4, "name": "BLEU"}, {"type": "chrf", "value": 0.54446, "name": "chr-F"}, {"type": "bleu", "value": 25.7, "name": "BLEU"}, {"type": "chrf", "value": 0.55237, "name": "chr-F"}, {"type": "bleu", "value": 36.9, "name": "BLEU"}, {"type": "chrf", "value": 0.62233, "name": "chr-F"}, {"type": "bleu", "value": 31.8, "name": "BLEU"}, {"type": "chrf", "value": 0.58235, "name": "chr-F"}, {"type": "bleu", "value": 24.3, "name": "BLEU"}, {"type": "chrf", "value": 0.52453, "name": "chr-F"}, {"type": "bleu", "value": 22.7, "name": "BLEU"}, {"type": "chrf", "value": 0.4893, "name": "chr-F"}, {"type": "bleu", "value": 26.2, "name": "BLEU"}, {"type": "chrf", "value": 0.52704, "name": "chr-F"}, {"type": "bleu", "value": 18.0, "name": "BLEU"}, {"type": "chrf", "value": 0.45387, "name": "chr-F"}, {"type": "bleu", "value": 18.6, "name": "BLEU"}, {"type": "chrf", "value": 0.47303, "name": "chr-F"}, {"type": "bleu", "value": 24.9, "name": "BLEU"}, {"type": "chrf", "value": 0.51381, "name": "chr-F"}, {"type": "bleu", "value": 21.6, "name": "BLEU"}, {"type": "chrf", "value": 0.48224, "name": "chr-F"}, {"type": "bleu", "value": 18.1, "name": "BLEU"}, {"type": "chrf", "value": 0.45786, "name": "chr-F"}, {"type": "bleu", "value": 28.9, "name": "BLEU"}, {"type": "chrf", "value": 0.55984, "name": "chr-F"}, {"type": "bleu", "value": 33.8, "name": "BLEU"}, {"type": "chrf", "value": 0.60102, "name": "chr-F"}, {"type": "bleu", "value": 23.4, "name": "BLEU"}, {"type": "chrf", "value": 0.52145, "name": "chr-F"}, {"type": "bleu", "value": 22.2, "name": "BLEU"}, {"type": "chrf", "value": 0.52619, "name": "chr-F"}, {"type": "bleu", "value": 32.2, "name": "BLEU"}, {"type": "chrf", "value": 0.58836, "name": "chr-F"}, {"type": "bleu", "value": 27.6, "name": "BLEU"}, {"type": "chrf", "value": 0.54845, "name": "chr-F"}, {"type": "bleu", "value": 21.8, "name": "BLEU"}, {"type": "chrf", "value": 0.50661, "name": "chr-F"}, {"type": "bleu", "value": 32.4, "name": "BLEU"}, {"type": "chrf", "value": 0.58542, "name": "chr-F"}, {"type": "bleu", "value": 39.3, "name": "BLEU"}, {"type": "chrf", "value": 0.63688, "name": "chr-F"}, {"type": "bleu", "value": 26.0, "name": "BLEU"}, {"type": "chrf", "value": 0.53989, "name": "chr-F"}, {"type": "bleu", "value": 25.9, "name": "BLEU"}, {"type": "chrf", "value": 0.55232, "name": "chr-F"}, {"type": "bleu", "value": 36.5, "name": "BLEU"}, {"type": "chrf", "value": 0.61882, "name": "chr-F"}, {"type": "bleu", "value": 31.0, "name": "BLEU"}, {"type": "chrf", "value": 0.57419, "name": "chr-F"}, {"type": "bleu", "value": 23.8, "name": "BLEU"}, {"type": "chrf", "value": 0.52175, "name": "chr-F"}]}, {"task": {"type": "translation", "name": "Translation dan-fra"}, "dataset": {"name": "tatoeba-test-v2021-08-07", "type": "tatoeba_mt", "args": "dan-fra"}, "metrics": [{"type": "bleu", "value": 63.8, "name": "BLEU"}, {"type": "chrf", "value": 0.76671, "name": "chr-F"}, {"type": "bleu", "value": 56.2, "name": "BLEU"}, {"type": "chrf", "value": 0.74658, "name": "chr-F"}, {"type": "bleu", "value": 57.8, "name": "BLEU"}, {"type": "chrf", "value": 0.74944, "name": "chr-F"}, {"type": "bleu", "value": 54.8, "name": "BLEU"}, {"type": "chrf", "value": 0.72328, "name": "chr-F"}, {"type": "bleu", "value": 51.0, "name": "BLEU"}, {"type": "chrf", "value": 0.69354, "name": "chr-F"}, {"type": "bleu", "value": 49.2, "name": "BLEU"}, {"type": "chrf", "value": 0.66008, "name": "chr-F"}, {"type": "bleu", "value": 54.4, "name": "BLEU"}, {"type": "chrf", "value": 0.70854, "name": "chr-F"}, {"type": "bleu", "value": 55.9, "name": "BLEU"}, {"type": "chrf", "value": 0.73672, "name": "chr-F"}, {"type": "bleu", "value": 59.2, "name": "BLEU"}, {"type": "chrf", "value": 0.73014, "name": "chr-F"}, {"type": "bleu", "value": 56.6, "name": "BLEU"}, {"type": "chrf", "value": 0.73211, "name": "chr-F"}, {"type": "bleu", "value": 48.7, "name": "BLEU"}, {"type": "chrf", "value": 0.68146, "name": "chr-F"}, {"type": "bleu", "value": 55.3, "name": "BLEU"}, {"type": "chrf", "value": 0.71373, "name": "chr-F"}]}]}]}	task	[ "TRANSLATION" ]	44,536
phospho-app/phospho-small-c99c7a6	phospho-app	feature-extraction	[ "transformers", "safetensors", "mpnet", "feature-extraction", "en", "license:apache-2.0", "endpoints_compatible", "region:us" ]	2024-04-30T08:38:30Z	2024-04-30T08:39:49+00:00	7	0	--- language: en license: apache-2.0 --- # phospho-small This is a SetFit model that can be used for Text Classification on CPU. The model has been trained using an efficient few-shot learning technique. ## Usage ```python from setfit import SetFitModel model = SetFitModel.from_pretrained("phospho-small-c99c7a6") outputs = model.predict(["This is a sentence to classify", "Another sentence"]) # tensor([1, 0]) ``` ## References This work was possible thanks to the SetFit library and the work of: Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren (2022). Efficient Few-Shot Learning Without Prompts. ArXiv: [https://doi.org/10.48550/arxiv.2209.11055](https://doi.org/10.48550/arxiv.2209.11055)	null	Non_BioNLP	# phospho-small This is a SetFit model that can be used for Text Classification on CPU. The model has been trained using an efficient few-shot learning technique. ## Usage ```python from setfit import SetFitModel model = SetFitModel.from_pretrained("phospho-small-c99c7a6") outputs = model.predict(["This is a sentence to classify", "Another sentence"]) # tensor([1, 0]) ``` ## References This work was possible thanks to the SetFit library and the work of: Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren (2022). Efficient Few-Shot Learning Without Prompts. ArXiv: [https://doi.org/10.48550/arxiv.2209.11055](https://doi.org/10.48550/arxiv.2209.11055)	{"language": "en", "license": "apache-2.0"}	task	[ "TEXT_CLASSIFICATION" ]	44,537
tosin/pcl_22	tosin	text2text-generation	[ "transformers", "pytorch", "t5", "text2text-generation", "text classification", "en", "dataset:PCL", "license:cc-by-4.0", "autotrain_compatible", "text-generation-inference", "region:us" ]	2022-03-02T23:29:05Z	2022-02-18T12:33:52+00:00	174	0	--- datasets: - PCL language: - en license: cc-by-4.0 metrics: - F1 tags: - text classification - transformers thumbnail: https://huggingface.co/front/thumbnails/dialogpt.png inference: false --- ## T5Base-PCL This is a fine-tuned model of T5 (base) on the patronizing and condenscending language (PCL) dataset by Pérez-Almendros et al (2020) used for Task 4 competition of SemEval-2022. It is intended to be used as a classification model for identifying PCL (0 - neg; 1 - pos). The task prefix we used for the T5 model is 'classification: '. The dataset it's trained on is limited in scope, as it covers only some news texts covering about 20 English-speaking countries. The macro F1 score achieved on the test set, based on the official evaluation, is 0.5452. More information about the original pre-trained model can be found [here](https://huggingface.co/t5-base) * Classification examples: \|Prediction \| Input \| \|---------\|------------\| \|0 \| selective kindness : in europe , some refugees are more equal than others \| \|1 \| he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty \| ### How to use ```python from transformers import T5ForConditionalGeneration, T5Tokenizer import torch model = T5ForConditionalGeneration.from_pretrained("tosin/pcl_22") tokenizer = T5Tokenizer.from_pretrained("t5-base") # use the source tokenizer because T5 finetuned tokenizer breaks tokenizer.pad_token = tokenizer.eos_token input_ids = tokenizer("he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty", padding=True, truncation=True, return_tensors='pt').input_ids outputs = model.generate(input_ids) pred = tokenizer.decode(outputs[0], skip_special_tokens=True) print(pred)	null	TBD	## T5Base-PCL This is a fine-tuned model of T5 (base) on the patronizing and condenscending language (PCL) dataset by Pérez-Almendros et al (2020) used for Task 4 competition of SemEval-2022. It is intended to be used as a classification model for identifying PCL (0 - neg; 1 - pos). The task prefix we used for the T5 model is 'classification: '. The dataset it's trained on is limited in scope, as it covers only some news texts covering about 20 English-speaking countries. The macro F1 score achieved on the test set, based on the official evaluation, is 0.5452. More information about the original pre-trained model can be found [here](https://huggingface.co/t5-base) * Classification examples: \|Prediction \| Input \| \|---------\|------------\| \|0 \| selective kindness : in europe , some refugees are more equal than others \| \|1 \| he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty \| ### How to use ```python from transformers import T5ForConditionalGeneration, T5Tokenizer import torch model = T5ForConditionalGeneration.from_pretrained("tosin/pcl_22") tokenizer = T5Tokenizer.from_pretrained("t5-base") # use the source tokenizer because T5 finetuned tokenizer breaks tokenizer.pad_token = tokenizer.eos_token input_ids = tokenizer("he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty", padding=True, truncation=True, return_tensors='pt').input_ids outputs = model.generate(input_ids) pred = tokenizer.decode(outputs[0], skip_special_tokens=True) print(pred)	{"datasets": ["PCL"], "language": ["en"], "license": "cc-by-4.0", "metrics": ["F1"], "tags": ["text classification", "transformers"], "thumbnail": "https://huggingface.co/front/thumbnails/dialogpt.png", "inference": false}	task	[ "TEXT_CLASSIFICATION" ]	44,538
gokuls/distilbert_sa_GLUE_Experiment_logit_kd_qnli	gokuls	text-classification	[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "en", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-01-29T22:09:31Z	2023-01-29T22:42:55+00:00	136	0	--- datasets: - glue language: - en license: apache-2.0 metrics: - accuracy tags: - generated_from_trainer model-index: - name: distilbert_sa_GLUE_Experiment_logit_kd_qnli results: - task: type: text-classification name: Text Classification dataset: name: GLUE QNLI type: glue config: qnli split: validation args: qnli metrics: - type: accuracy value: 0.5905180303862346 name: Accuracy --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert_sa_GLUE_Experiment_logit_kd_qnli This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the GLUE QNLI dataset. It achieves the following results on the evaluation set: - Loss: 0.3865 - Accuracy: 0.5905 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - distributed_type: multi-GPU - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 50 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.4069 \| 1.0 \| 410 \| 0.3914 \| 0.5680 \| \| 0.3877 \| 2.0 \| 820 \| 0.3865 \| 0.5905 \| \| 0.3741 \| 3.0 \| 1230 \| 0.3917 \| 0.5971 \| \| 0.3604 \| 4.0 \| 1640 \| 0.3893 \| 0.5929 \| \| 0.3432 \| 5.0 \| 2050 \| 0.3908 \| 0.5922 \| \| 0.3194 \| 6.0 \| 2460 \| 0.4251 \| 0.5861 \| \| 0.2938 \| 7.0 \| 2870 \| 0.4414 \| 0.5940 \| ### Framework versions - Transformers 4.26.0 - Pytorch 1.14.0a0+410ce96 - Datasets 2.9.0 - Tokenizers 0.13.2	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert_sa_GLUE_Experiment_logit_kd_qnli This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the GLUE QNLI dataset. It achieves the following results on the evaluation set: - Loss: 0.3865 - Accuracy: 0.5905 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - distributed_type: multi-GPU - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 50 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.4069 \| 1.0 \| 410 \| 0.3914 \| 0.5680 \| \| 0.3877 \| 2.0 \| 820 \| 0.3865 \| 0.5905 \| \| 0.3741 \| 3.0 \| 1230 \| 0.3917 \| 0.5971 \| \| 0.3604 \| 4.0 \| 1640 \| 0.3893 \| 0.5929 \| \| 0.3432 \| 5.0 \| 2050 \| 0.3908 \| 0.5922 \| \| 0.3194 \| 6.0 \| 2460 \| 0.4251 \| 0.5861 \| \| 0.2938 \| 7.0 \| 2870 \| 0.4414 \| 0.5940 \| ### Framework versions - Transformers 4.26.0 - Pytorch 1.14.0a0+410ce96 - Datasets 2.9.0 - Tokenizers 0.13.2	{"datasets": ["glue"], "language": ["en"], "license": "apache-2.0", "metrics": ["accuracy"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert_sa_GLUE_Experiment_logit_kd_qnli", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "GLUE QNLI", "type": "glue", "config": "qnli", "split": "validation", "args": "qnli"}, "metrics": [{"type": "accuracy", "value": 0.5905180303862346, "name": "Accuracy"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,539
Alphatao/f2b0fa58-8c68-4ee4-8ff5-65c996bc267d	Alphatao	null	[ "peft", "safetensors", "llama", "axolotl", "generated_from_trainer", "base_model:llamafactory/tiny-random-Llama-3", "base_model:adapter:llamafactory/tiny-random-Llama-3", "license:apache-2.0", "region:us" ]	2025-03-10T07:37:37Z	2025-03-10T08:02:44+00:00	8	0	--- base_model: llamafactory/tiny-random-Llama-3 library_name: peft license: apache-2.0 tags: - axolotl - generated_from_trainer model-index: - name: f2b0fa58-8c68-4ee4-8ff5-65c996bc267d results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: llamafactory/tiny-random-Llama-3 bf16: true chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 420d808b1b64ffc0_train_data.json ds_type: json format: custom path: /workspace/input_data/420d808b1b64ffc0_train_data.json type: field_instruction: source field_output: good-translation format: '{instruction}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null device_map: ? '' : 0,1,2,3,4,5,6,7 early_stopping_patience: 2 eval_max_new_tokens: 128 eval_steps: 100 eval_table_size: null flash_attention: true gradient_accumulation_steps: 8 gradient_checkpointing: true group_by_length: false hub_model_id: Alphatao/f2b0fa58-8c68-4ee4-8ff5-65c996bc267d hub_repo: null hub_strategy: null hub_token: null learning_rate: 0.0002 load_best_model_at_end: true load_in_4bit: false load_in_8bit: false local_rank: null logging_steps: 1 lora_alpha: 128 lora_dropout: 0.1 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 64 lora_target_linear: true lora_target_modules: - q_proj - k_proj - v_proj - o_proj lr_scheduler: cosine max_grad_norm: 1.0 max_steps: 5376 micro_batch_size: 4 mlflow_experiment_name: /tmp/420d808b1b64ffc0_train_data.json model_type: AutoModelForCausalLM num_epochs: 2 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false save_steps: 100 sequence_len: 2048 special_tokens: pad_token: <\|eot_id\|> strict: false tf32: true tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.04 wandb_entity: null wandb_mode: online wandb_name: 70c61e5c-5f68-4639-804d-89212f8c7742 wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 70c61e5c-5f68-4639-804d-89212f8c7742 warmup_steps: 10 weight_decay: 0.0 xformers_attention: null ``` </details><br> # f2b0fa58-8c68-4ee4-8ff5-65c996bc267d This model is a fine-tuned version of [llamafactory/tiny-random-Llama-3](https://huggingface.co/llamafactory/tiny-random-Llama-3) on the None dataset. It achieves the following results on the evaluation set: - Loss: 11.7233 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 32 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 10 - training_steps: 2123 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:------:\|:----:\|:---------------:\| \| 11.7614 \| 0.0009 \| 1 \| 11.7621 \| \| 11.7349 \| 0.0942 \| 100 \| 11.7364 \| \| 11.7389 \| 0.1885 \| 200 \| 11.7346 \| \| 11.732 \| 0.2827 \| 300 \| 11.7321 \| \| 11.7438 \| 0.3769 \| 400 \| 11.7298 \| \| 11.7335 \| 0.4711 \| 500 \| 11.7285 \| \| 11.7324 \| 0.5654 \| 600 \| 11.7275 \| \| 11.7331 \| 0.6596 \| 700 \| 11.7265 \| \| 11.7274 \| 0.7538 \| 800 \| 11.7258 \| \| 11.7295 \| 0.8481 \| 900 \| 11.7255 \| \| 11.731 \| 0.9423 \| 1000 \| 11.7251 \| \| 13.0224 \| 1.0365 \| 1100 \| 11.7249 \| \| 11.6643 \| 1.1307 \| 1200 \| 11.7242 \| \| 11.7749 \| 1.2250 \| 1300 \| 11.7240 \| \| 11.8175 \| 1.3192 \| 1400 \| 11.7238 \| \| 10.3085 \| 1.4134 \| 1500 \| 11.7236 \| \| 12.9048 \| 1.5077 \| 1600 \| 11.7235 \| \| 10.6263 \| 1.6019 \| 1700 \| 11.7234 \| \| 10.6834 \| 1.6961 \| 1800 \| 11.7233 \| \| 11.5317 \| 1.7903 \| 1900 \| 11.7233 \| \| 12.3242 \| 1.8846 \| 2000 \| 11.7233 \| \| 10.9003 \| 1.9788 \| 2100 \| 11.7233 \| ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: llamafactory/tiny-random-Llama-3 bf16: true chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 420d808b1b64ffc0_train_data.json ds_type: json format: custom path: /workspace/input_data/420d808b1b64ffc0_train_data.json type: field_instruction: source field_output: good-translation format: '{instruction}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null device_map: ? '' : 0,1,2,3,4,5,6,7 early_stopping_patience: 2 eval_max_new_tokens: 128 eval_steps: 100 eval_table_size: null flash_attention: true gradient_accumulation_steps: 8 gradient_checkpointing: true group_by_length: false hub_model_id: Alphatao/f2b0fa58-8c68-4ee4-8ff5-65c996bc267d hub_repo: null hub_strategy: null hub_token: null learning_rate: 0.0002 load_best_model_at_end: true load_in_4bit: false load_in_8bit: false local_rank: null logging_steps: 1 lora_alpha: 128 lora_dropout: 0.1 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 64 lora_target_linear: true lora_target_modules: - q_proj - k_proj - v_proj - o_proj lr_scheduler: cosine max_grad_norm: 1.0 max_steps: 5376 micro_batch_size: 4 mlflow_experiment_name: /tmp/420d808b1b64ffc0_train_data.json model_type: AutoModelForCausalLM num_epochs: 2 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false save_steps: 100 sequence_len: 2048 special_tokens: pad_token: <\|eot_id\|> strict: false tf32: true tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.04 wandb_entity: null wandb_mode: online wandb_name: 70c61e5c-5f68-4639-804d-89212f8c7742 wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 70c61e5c-5f68-4639-804d-89212f8c7742 warmup_steps: 10 weight_decay: 0.0 xformers_attention: null ``` </details><br> # f2b0fa58-8c68-4ee4-8ff5-65c996bc267d This model is a fine-tuned version of [llamafactory/tiny-random-Llama-3](https://huggingface.co/llamafactory/tiny-random-Llama-3) on the None dataset. It achieves the following results on the evaluation set: - Loss: 11.7233 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 32 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 10 - training_steps: 2123 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:------:\|:----:\|:---------------:\| \| 11.7614 \| 0.0009 \| 1 \| 11.7621 \| \| 11.7349 \| 0.0942 \| 100 \| 11.7364 \| \| 11.7389 \| 0.1885 \| 200 \| 11.7346 \| \| 11.732 \| 0.2827 \| 300 \| 11.7321 \| \| 11.7438 \| 0.3769 \| 400 \| 11.7298 \| \| 11.7335 \| 0.4711 \| 500 \| 11.7285 \| \| 11.7324 \| 0.5654 \| 600 \| 11.7275 \| \| 11.7331 \| 0.6596 \| 700 \| 11.7265 \| \| 11.7274 \| 0.7538 \| 800 \| 11.7258 \| \| 11.7295 \| 0.8481 \| 900 \| 11.7255 \| \| 11.731 \| 0.9423 \| 1000 \| 11.7251 \| \| 13.0224 \| 1.0365 \| 1100 \| 11.7249 \| \| 11.6643 \| 1.1307 \| 1200 \| 11.7242 \| \| 11.7749 \| 1.2250 \| 1300 \| 11.7240 \| \| 11.8175 \| 1.3192 \| 1400 \| 11.7238 \| \| 10.3085 \| 1.4134 \| 1500 \| 11.7236 \| \| 12.9048 \| 1.5077 \| 1600 \| 11.7235 \| \| 10.6263 \| 1.6019 \| 1700 \| 11.7234 \| \| 10.6834 \| 1.6961 \| 1800 \| 11.7233 \| \| 11.5317 \| 1.7903 \| 1900 \| 11.7233 \| \| 12.3242 \| 1.8846 \| 2000 \| 11.7233 \| \| 10.9003 \| 1.9788 \| 2100 \| 11.7233 \| ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1	{"base_model": "llamafactory/tiny-random-Llama-3", "library_name": "peft", "license": "apache-2.0", "tags": ["axolotl", "generated_from_trainer"], "model-index": [{"name": "f2b0fa58-8c68-4ee4-8ff5-65c996bc267d", "results": []}]}	task	[ "TRANSLATION" ]	44,541
aimped/nlp-health-translation-base-en-fr	aimped	translation	[ "transformers", "pytorch", "marian", "text2text-generation", "medical", "translation", "medical translation", "en", "fr", "dataset:aimped/medical-translation-test-set", "arxiv:2407.12126", "license:cc-by-nc-4.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-10-26T08:55:01Z	2024-08-29T10:20:55+00:00	0	1	--- datasets: - aimped/medical-translation-test-set language: - en - fr license: cc-by-nc-4.0 metrics: - bleu pipeline_tag: translation tags: - medical - translation - medical translation extra_gated_heading: Access aimped/nlp-health-translation-base-en-fr on Hugging Face extra_gated_description: This is a form to enable access to this model on Hugging Face after you have been granted access from the Aimped. Please visit the [Aimped website](https://aimped.ai/) to Sign Up and accept our Terms of Use and Privacy Policy before submitting this form. Requests will be processed in 1-2 days. extra_gated_prompt: 'Your Hugging Face account email address MUST match the email you provide on the Aimped website or your request will not be approved.' extra_gated_button_content: Submit extra_gated_fields: ? I agree to share my name, email address, and username with Aimped and confirm that I have already been granted download access on the Aimped website : checkbox widget: - text: 'Objective: Physical traumas are one of the important causes of mortality and morbidity in childhood. Permanent disabilities resulting from traumas constitute significant losses for the individual and society.' - text: Evidence is reported that a variety of chronic respiratory diseases, particularly COPD, asthma, bronchiectasis, lung cancer, interstitial lung diseases, and sarcoidosis, are significantly associated with poor clinical outcomes of COVID-19. --- <p> <p align="center"> <img src="https://raw.githubusercontent.com/ai-amplified/models/main/media/AimpedLogoDark.svg" alt="aimped logo" width="50%" height="50%"/> </p> ### Description of the Model <p> Paper: <a href="https://arxiv.org/abs/2407.12126" style="text-decoration: underline; color: blue;">LLMs-in-the-loop Part-1: Expert Small AI Models for Bio-Medical Text Translation</a> </p> </p> <p style="margin-bottom: 0in; text-align: justify; line-height: 1.3;"><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">The Medical Translation AI model represents a specialized language model, trained for the accurate translations of medical documents from English to French. Its primary objective is to provide healthcare professionals, researchers, and individuals within the medical field with a reliable tool for the precise translation of a wide spectrum of medical documents.            </span></p> <p style="margin-bottom: 0in; text-align: justify; line-height: 1.3;"> <span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">The development of this model entailed the utilization of the <a href="https://github.com/Helsinki-NLP/OPUS-MT-train/tree/master/models/en-fr" style="text-decoration: underline; color: blue;">Hensinki/MarianMT</a> neural translation architecture, which required 2+ days of intensive training using A100 (24G RAM) GPU. To create an exceptionally high-quality corpus for training the translation model, we combined both publicly available and proprietary datasets. These datasets were further enriched by meticulously curated text collected from online sources. In addition, the inclusion of clinical and discharge reports from diverse healthcare institutions enhanced the dataset's depth and diversity. This meticulous curation process plays a pivotal role in ensuring the model's ability to generate accurate translations tailored specifically to the medical domain, meeting the stringent standards expected by our users.<br><br>The versatility of the Medical Translation AI model extends to the translation of a wide array of healthcare-related documents, encompassing medical reports, patient records, medication instructions, research manuscripts, clinical trial documents, and more. By harnessing the capabilities of this model, users can efficiently and dependably obtain translations, thereby streamlining and expediting the often complex task of language translation within the medical field.</span> </p> <p style="margin-bottom: 0in; text-align: justify; line-height: 1.3;"><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">The model we have developed outperforms leading translation companies like Google, Helsinki-Opus/MarianMT, and DeepL when compared against our meticulously curated proprietary test data set. </span></p> <p style="line-height: 1.3; margin-bottom: 0in; text-align: justify;"><br></p> <table style="border-collapse: collapse; width: 605px; height: 117px; border: 1px lightgray;"> <tbody> <tr> <td style="width: 19.5041%; border: 1px lightgray;"><br></td> <td style="width: 20.6612%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>ROUGE</strong></td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>BLEU</strong></td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>METEOR</strong></td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>BERT</strong></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>Aimped</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.85</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.62</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.83</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.95</span></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>Google</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.84</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.61</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.82</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.95</span></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>DeepL</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.81</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.57</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.78</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.94</span></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>Opus/MarianMT</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.80</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.52</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.76</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.93</span></td> </tr> </tbody> </table> <p></p> ## Why should you use Aimped API? To get started, you can easily use our open-source version of the models for research purposes. However, the models provided through the Aimped API are trained on new data every three months. This ensures that the models understand ongoing healthcare developments in the world and can identify the most relevant medical terminology without a knowledge cutoff. In addition, we implement post/pre processing steps to improve the translation quality. Naturally, our quality control ensures that the models' performance always remains at least similar to previous versions. ## How to Use: To get the right results, use this function. - Install requirements ```python !pip install transformers !pip install sentencepiece !pip install aimped import nltk nltk.download('punkt') ``` - import libraries ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline from aimped.nlp.translation import text_translate import torch device = "cuda" if torch.cuda.is_available() else "cpu" ``` - load model ```python model_path = "aimped/nlp-health-translation-base-en-fr" tokenizer = AutoTokenizer.from_pretrained(model_path) model = AutoModelForSeq2SeqLM.from_pretrained(model_path) ``` ```python translater = pipeline( task="translation_en_to_fr", model=model, tokenizer=tokenizer, device= device, max_length=512, num_beams=7, early_stopping=False, num_return_sequences=1, do_sample=False, ) ``` - Use Model: ```python sentence = "Conclusion: According to our findings, the most common causes of major injuries in childhood are falls and home accidents." translated_text = text_translate([sentence],source_lang="en", pipeline=translater) ``` ## Test Set <p><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">Trainin data: Public and in-house datasets.</span></p> <p><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">Test data: Public and in-house datasets which is available <a href="https://github.com/ai-amplified/models/tree/main/medical_translation/test_data/en-fr pairs">here</a>.</span></p><br class="Apple-interchange-newline">	null	BioNLP	<p> <p align="center"> <img src="https://raw.githubusercontent.com/ai-amplified/models/main/media/AimpedLogoDark.svg" alt="aimped logo" width="50%" height="50%"/> </p> ### Description of the Model <p> Paper: <a href="https://arxiv.org/abs/2407.12126" style="text-decoration: underline; color: blue;">LLMs-in-the-loop Part-1: Expert Small AI Models for Bio-Medical Text Translation</a> </p> </p> <p style="margin-bottom: 0in; text-align: justify; line-height: 1.3;"><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">The Medical Translation AI model represents a specialized language model, trained for the accurate translations of medical documents from English to French. Its primary objective is to provide healthcare professionals, researchers, and individuals within the medical field with a reliable tool for the precise translation of a wide spectrum of medical documents.            </span></p> <p style="margin-bottom: 0in; text-align: justify; line-height: 1.3;"> <span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">The development of this model entailed the utilization of the <a href="https://github.com/Helsinki-NLP/OPUS-MT-train/tree/master/models/en-fr" style="text-decoration: underline; color: blue;">Hensinki/MarianMT</a> neural translation architecture, which required 2+ days of intensive training using A100 (24G RAM) GPU. To create an exceptionally high-quality corpus for training the translation model, we combined both publicly available and proprietary datasets. These datasets were further enriched by meticulously curated text collected from online sources. In addition, the inclusion of clinical and discharge reports from diverse healthcare institutions enhanced the dataset's depth and diversity. This meticulous curation process plays a pivotal role in ensuring the model's ability to generate accurate translations tailored specifically to the medical domain, meeting the stringent standards expected by our users.<br><br>The versatility of the Medical Translation AI model extends to the translation of a wide array of healthcare-related documents, encompassing medical reports, patient records, medication instructions, research manuscripts, clinical trial documents, and more. By harnessing the capabilities of this model, users can efficiently and dependably obtain translations, thereby streamlining and expediting the often complex task of language translation within the medical field.</span> </p> <p style="margin-bottom: 0in; text-align: justify; line-height: 1.3;"><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">The model we have developed outperforms leading translation companies like Google, Helsinki-Opus/MarianMT, and DeepL when compared against our meticulously curated proprietary test data set. </span></p> <p style="line-height: 1.3; margin-bottom: 0in; text-align: justify;"><br></p> <table style="border-collapse: collapse; width: 605px; height: 117px; border: 1px lightgray;"> <tbody> <tr> <td style="width: 19.5041%; border: 1px lightgray;"><br></td> <td style="width: 20.6612%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>ROUGE</strong></td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>BLEU</strong></td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>METEOR</strong></td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;"><strong>BERT</strong></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>Aimped</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.85</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.62</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.83</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.95</span></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>Google</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.84</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.61</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.82</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.95</span></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>DeepL</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.81</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.57</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.78</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.94</span></td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>Opus/MarianMT</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.80</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.52</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.76</span></td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;"><span>0.93</span></td> </tr> </tbody> </table> <p></p> ## Why should you use Aimped API? To get started, you can easily use our open-source version of the models for research purposes. However, the models provided through the Aimped API are trained on new data every three months. This ensures that the models understand ongoing healthcare developments in the world and can identify the most relevant medical terminology without a knowledge cutoff. In addition, we implement post/pre processing steps to improve the translation quality. Naturally, our quality control ensures that the models' performance always remains at least similar to previous versions. ## How to Use: To get the right results, use this function. - Install requirements ```python !pip install transformers !pip install sentencepiece !pip install aimped import nltk nltk.download('punkt') ``` - import libraries ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline from aimped.nlp.translation import text_translate import torch device = "cuda" if torch.cuda.is_available() else "cpu" ``` - load model ```python model_path = "aimped/nlp-health-translation-base-en-fr" tokenizer = AutoTokenizer.from_pretrained(model_path) model = AutoModelForSeq2SeqLM.from_pretrained(model_path) ``` ```python translater = pipeline( task="translation_en_to_fr", model=model, tokenizer=tokenizer, device= device, max_length=512, num_beams=7, early_stopping=False, num_return_sequences=1, do_sample=False, ) ``` - Use Model: ```python sentence = "Conclusion: According to our findings, the most common causes of major injuries in childhood are falls and home accidents." translated_text = text_translate([sentence],source_lang="en", pipeline=translater) ``` ## Test Set <p><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">Trainin data: Public and in-house datasets.</span></p> <p><span style="font-family: "IBM Plex Sans", sans-serif; font-size: 16px;">Test data: Public and in-house datasets which is available <a href="https://github.com/ai-amplified/models/tree/main/medical_translation/test_data/en-fr pairs">here</a>.</span></p><br class="Apple-interchange-newline">	{"datasets": ["aimped/medical-translation-test-set"], "language": ["en", "fr"], "license": "cc-by-nc-4.0", "metrics": ["bleu"], "pipeline_tag": "translation", "tags": ["medical", "translation", "medical translation"], "extra_gated_heading": "Access aimped/nlp-health-translation-base-en-fr on Hugging Face", "extra_gated_description": "This is a form to enable access to this model on Hugging Face after you have been granted access from the Aimped. Please visit the [Aimped website](https://aimped.ai/) to Sign Up and accept our Terms of Use and Privacy Policy before submitting this form. Requests will be processed in 1-2 days.", "extra_gated_prompt": "Your Hugging Face account email address MUST match the email you provide on the Aimped website or your request will not be approved.", "extra_gated_button_content": "Submit", "extra_gated_fields": {"I agree to share my name, email address, and username with Aimped and confirm that I have already been granted download access on the Aimped website": "checkbox"}, "widget": [{"text": "Objective: Physical traumas are one of the important causes of mortality and morbidity in childhood. Permanent disabilities resulting from traumas constitute significant losses for the individual and society."}, {"text": "Evidence is reported that a variety of chronic respiratory diseases, particularly COPD, asthma, bronchiectasis, lung cancer, interstitial lung diseases, and sarcoidosis, are significantly associated with poor clinical outcomes of COVID-19."}]}	task	[ "TRANSLATION" ]	44,542
PardisSzah/Persian_NER_parsbert	PardisSzah	token-classification	[ "transformers", "safetensors", "bert", "token-classification", "fa", "license:mit", "autotrain_compatible", "region:us" ]	2024-02-16T19:26:01Z	2024-10-21T12:57:51+00:00	22	0	--- language: fa license: mit pipeline_tag: token-classification inference: false --- # NER_ARMAN_parsbert This model is fine-tuned for Named Entity Recognition task. It has been fine-tuned on ARMAN Dataset, using the pretrained model [bert-base-parsbert-ner-uncased](https://huggingface.co/HooshvareLab/bert-base-parsbert-ner-uncased). ## Usage ```python def predict(input_text): nlp = pipeline("ner", model="PardisSzah/Persian_NER_parsbert") output_predictions = [] for sequence in input_text: result = nlp(sequence) output_predictions.append(result) return output_predictions text = [ "علی اکبری در روز شنبه به دیدن مادرش مریم حسنی رفت و بعد به بیمارستان ارتش سر زد" ] output = predict(text) print(output) # output: [[{'entity': 'B-person', 'score': 0.9998951, 'index': 1, 'word': 'علی', 'start': 0, 'end': 3}, {'entity': 'I-person', 'score': 0.9999027, 'index': 2, 'word': 'اکبری', 'start': 4, 'end': 9}, {'entity': 'B-person', 'score': 0.9998709, 'index': 9, 'word': 'مریم', 'start': 36, 'end': 40}, {'entity': 'I-person', 'score': 0.9996691, 'index': 10, 'word': 'حسنی', 'start': 41, 'end': 45}, {'entity': 'B-facility', 'score': 0.9561743, 'index': 15, 'word': 'بیمارستان', 'start': 59, 'end': 68}, {'entity': 'I-facility', 'score': 0.9976502, 'index': 16, 'word': 'ارتش', 'start': 69, 'end': 73}]]	null	Non_BioNLP	# NER_ARMAN_parsbert This model is fine-tuned for Named Entity Recognition task. It has been fine-tuned on ARMAN Dataset, using the pretrained model [bert-base-parsbert-ner-uncased](https://huggingface.co/HooshvareLab/bert-base-parsbert-ner-uncased). ## Usage ```python def predict(input_text): nlp = pipeline("ner", model="PardisSzah/Persian_NER_parsbert") output_predictions = [] for sequence in input_text: result = nlp(sequence) output_predictions.append(result) return output_predictions text = [ "علی اکبری در روز شنبه به دیدن مادرش مریم حسنی رفت و بعد به بیمارستان ارتش سر زد" ] output = predict(text) print(output) # output: [[{'entity': 'B-person', 'score': 0.9998951, 'index': 1, 'word': 'علی', 'start': 0, 'end': 3}, {'entity': 'I-person', 'score': 0.9999027, 'index': 2, 'word': 'اکبری', 'start': 4, 'end': 9}, {'entity': 'B-person', 'score': 0.9998709, 'index': 9, 'word': 'مریم', 'start': 36, 'end': 40}, {'entity': 'I-person', 'score': 0.9996691, 'index': 10, 'word': 'حسنی', 'start': 41, 'end': 45}, {'entity': 'B-facility', 'score': 0.9561743, 'index': 15, 'word': 'بیمارستان', 'start': 59, 'end': 68}, {'entity': 'I-facility', 'score': 0.9976502, 'index': 16, 'word': 'ارتش', 'start': 69, 'end': 73}]]	{"language": "fa", "license": "mit", "pipeline_tag": "token-classification", "inference": false}	task	[ "NAMED_ENTITY_RECOGNITION" ]	44,543
broyal/gemma-7b-AWQ	broyal	text-generation	[ "transformers", "safetensors", "gemma", "text-generation", "arxiv:2305.14314", "arxiv:2312.11805", "arxiv:2009.03300", "arxiv:1905.07830", "arxiv:1911.11641", "arxiv:1904.09728", "arxiv:1905.10044", "arxiv:1907.10641", "arxiv:1811.00937", "arxiv:1809.02789", "arxiv:1911.01547", "arxiv:1705.03551", "arxiv:2107.03374", "arxiv:2108.07732", "arxiv:2110.14168", "arxiv:2304.06364", "arxiv:2206.04615", "arxiv:1804.06876", "arxiv:2110.08193", "arxiv:2009.11462", "arxiv:2101.11718", "arxiv:1804.09301", "arxiv:2109.07958", "arxiv:2203.09509", "license:other", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "4-bit", "awq", "region:us" ]	2024-07-19T13:44:20Z	2024-07-19T15:18:23+00:00	77	0	--- library_name: transformers license: other license_name: gemma-terms-of-use license_link: https://ai.google.dev/gemma/terms tags: [] extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged-in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- AWQ quantized version of gemma-7b model. --- # Gemma Model Card Model Page: [Gemma](https://ai.google.dev/gemma/docs) This model card corresponds to the 7B base version of the Gemma model. You can also visit the model card of the [2B base model](https://huggingface.co/google/gemma-2b), [7B instruct model](https://huggingface.co/google/gemma-7b-it), and [2B instruct model](https://huggingface.co/google/gemma-2b-it). Resources and Technical Documentation: * [Responsible Generative AI Toolkit](https://ai.google.dev/responsible) * [Gemma on Kaggle](https://www.kaggle.com/models/google/gemma) * [Gemma on Vertex Model Garden](https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335?version=gemma-7b-gg-hf) Terms of Use: [Terms](https://www.kaggle.com/models/google/gemma/license/consent) Authors: Google ## Model Information Summary description and brief definition of inputs and outputs. ### Description Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. They are text-to-text, decoder-only large language models, available in English, with open weights, pre-trained variants, and instruction-tuned variants. Gemma models are well-suited for a variety of text generation tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as a laptop, desktop or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. ### Usage Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase. #### Fine-tuning examples You can find fine-tuning notebooks under the [`examples/` directory](https://huggingface.co/google/gemma-7b/tree/main/examples). We provide: * A script to perform Supervised Fine-Tuning (SFT) on UltraChat dataset using [QLoRA](https://huggingface.co/papers/2305.14314) * A script to perform SFT using FSDP on TPU devices * A notebook that you can run on a free-tier Google Colab instance to perform SFT on English quotes dataset. You can also find the copy of the notebook [here](https://github.com/huggingface/notebooks/blob/main/peft/gemma_7b_english_quotes.ipynb). #### Running the model on a CPU ```python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt") outputs = model.generate(input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a GPU using different precisions * _Using `torch.float16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.float16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` _Using `torch.bfloat16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.bfloat16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` _Using 8-bit precision (int8)_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` _Using 4-bit precision_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` #### Other optimizations _Flash Attention 2_ First make sure to install `flash-attn` in your environment `pip install flash-attn` ```diff model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, + attn_implementation="flash_attention_2" ).to(0) ``` ### Inputs and outputs * Input: Text string, such as a question, a prompt, or a document to be summarized. * Output: Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ## Model Data Data used for model training and how the data was processed. ### Training Dataset These models were trained on a dataset of text data that includes a wide variety of sources, totaling 6 trillion tokens. Here are the key components: * Web Documents: A diverse collection of web text ensures the model is exposed to a broad range of linguistic styles, topics, and vocabulary. Primarily English-language content. * Code: Exposing the model to code helps it to learn the syntax and patterns of programming languages, which improves its ability to generate code or understand code-related questions. * Mathematics: Training on mathematical text helps the model learn logical reasoning, symbolic representation, and to address mathematical queries. The combination of these diverse data sources is crucial for training a powerful language model that can handle a wide variety of different tasks and text formats. ### Data Preprocessing Here are the key data cleaning and filtering methods applied to the training data: * CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering was applied at multiple stages in the data preparation process to ensure the exclusion of harmful and illegal content * Sensitive Data Filtering: As part of making Gemma pre-trained models safe and reliable, automated techniques were used to filter out certain personal information and other sensitive data from training sets. * Additional methods: Filtering based on content quality and safely in line with [our policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11). ## Implementation Information Details about the model internals. ### Hardware Gemma was trained using the latest generation of [Tensor Processing Unit (TPU)](https://cloud.google.com/tpu/docs/intro-to-tpu) hardware (TPUv5e). Training large language models requires significant computational power. TPUs, designed specifically for matrix operations common in machine learning, offer several advantages in this domain: * Performance: TPUs are specifically designed to handle the massive computations involved in training LLMs. They can speed up training considerably compared to CPUs. * Memory: TPUs often come with large amounts of high-bandwidth memory, allowing for the handling of large models and batch sizes during training. This can lead to better model quality. * Scalability: TPU Pods (large clusters of TPUs) provide a scalable solution for handling the growing complexity of large foundation models. You can distribute training across multiple TPU devices for faster and more efficient processing. * Cost-effectiveness: In many scenarios, TPUs can provide a more cost-effective solution for training large models compared to CPU-based infrastructure, especially when considering the time and resources saved due to faster training. * These advantages are aligned with [Google's commitments to operate sustainably](https://sustainability.google/operating-sustainably/). ### Software Training was done using [JAX](https://github.com/google/jax) and [ML Pathways](https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture). JAX allows researchers to take advantage of the latest generation of hardware, including TPUs, for faster and more efficient training of large models. ML Pathways is Google's latest effort to build artificially intelligent systems capable of generalizing across multiple tasks. This is specially suitable for [foundation models](https://ai.google/discover/foundation-models/), including large language models like these ones. Together, JAX and ML Pathways are used as described in the [paper about the Gemini family of models](https://arxiv.org/abs/2312.11805); "the 'single controller' programming model of Jax and Pathways allows a single Python process to orchestrate the entire training run, dramatically simplifying the development workflow." ## Evaluation Model evaluation metrics and results. ### Benchmark Results These models were evaluated against a large collection of different datasets and metrics to cover different aspects of text generation: \| Benchmark \| Metric \| 2B Params \| 7B Params \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| \| [MMLU](https://arxiv.org/abs/2009.03300) \| 5-shot, top-1 \| 42.3 \| 64.3 \| \| [HellaSwag](https://arxiv.org/abs/1905.07830) \| 0-shot \|71.4 \| 81.2 \| \| [PIQA](https://arxiv.org/abs/1911.11641) \| 0-shot \| 77.3 \| 81.2 \| \| [SocialIQA](https://arxiv.org/abs/1904.09728) \| 0-shot \| 59.7 \| 51.8 \| \| [BooIQ](https://arxiv.org/abs/1905.10044) \| 0-shot \| 69.4 \| 83.2 \| \| [WinoGrande](https://arxiv.org/abs/1907.10641) \| partial score \| 65.4 \| 72.3 \| \| [CommonsenseQA](https://arxiv.org/abs/1811.00937) \| 7-shot \| 65.3 \| 71.3 \| \| [OpenBookQA](https://arxiv.org/abs/1809.02789) \| \| 47.8 \| 52.8 \| \| [ARC-e](https://arxiv.org/abs/1911.01547) \| \| 73.2 \| 81.5 \| \| [ARC-c](https://arxiv.org/abs/1911.01547) \| \| 42.1 \| 53.2 \| \| [TriviaQA](https://arxiv.org/abs/1705.03551) \| 5-shot \| 53.2 \| 63.4 \| \| [Natural Questions](https://github.com/google-research-datasets/natural-questions) \| 5-shot \| - \| 23 \| \| [HumanEval](https://arxiv.org/abs/2107.03374) \| pass@1 \| 22.0 \| 32.3 \| \| [MBPP](https://arxiv.org/abs/2108.07732) \| 3-shot \| 29.2 \| 44.4 \| \| [GSM8K](https://arxiv.org/abs/2110.14168) \| maj@1 \| 17.7 \| 46.4 \| \| [MATH](https://arxiv.org/abs/2108.07732) \| 4-shot \| 11.8 \| 24.3 \| \| [AGIEval](https://arxiv.org/abs/2304.06364) \| \| 24.2 \| 41.7 \| \| [BIG-Bench](https://arxiv.org/abs/2206.04615) \| \| 35.2 \| 55.1 \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| \| Average \| \| 54.0 \| 56.4 \| ## Ethics and Safety Ethics and safety evaluation approach and results. ### Evaluation Approach Our evaluation methods include structured evaluations and internal red-teaming testing of relevant content policies. Red-teaming was conducted by a number of different teams, each with different goals and human evaluation metrics. These models were evaluated against a number of different categories relevant to ethics and safety, including: * Text-to-Text Content Safety: Human evaluation on prompts covering safety policies including child sexual abuse and exploitation, harassment, violence and gore, and hate speech. * Text-to-Text Representational Harms: Benchmark against relevant academic datasets such as [WinoBias](https://arxiv.org/abs/1804.06876) and [BBQ Dataset](https://arxiv.org/abs/2110.08193v2). * Memorization: Automated evaluation of memorization of training data, including the risk of personally identifiable information exposure. * Large-scale harm: Tests for "dangerous capabilities," such as chemical, biological, radiological, and nuclear (CBRN) risks. ### Evaluation Results The results of ethics and safety evaluations are within acceptable thresholds for meeting [internal policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11) for categories such as child safety, content safety, representational harms, memorization, large-scale harms. On top of robust internal evaluations, the results of well known safety benchmarks like BBQ, BOLD, Winogender, Winobias, RealToxicity, and TruthfulQA are shown here. \| Benchmark \| Metric \| 2B Params \| 7B Params \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| \| [RealToxicity](https://arxiv.org/abs/2009.11462) \| average \| 6.86 \| 7.90 \| \| [BOLD](https://arxiv.org/abs/2101.11718) \| \| 45.57 \| 49.08 \| \| [CrowS-Pairs](https://aclanthology.org/2020.emnlp-main.154/) \| top-1 \| 45.82 \| 51.33 \| \| [BBQ Ambig](https://arxiv.org/abs/2110.08193v2) \| 1-shot, top-1 \| 62.58 \| 92.54 \| \| [BBQ Disambig](https://arxiv.org/abs/2110.08193v2) \| top-1 \| 54.62 \| 71.99 \| \| [Winogender](https://arxiv.org/abs/1804.09301) \| top-1 \| 51.25 \| 54.17 \| \| [TruthfulQA](https://arxiv.org/abs/2109.07958) \| \| 44.84 \| 31.81 \| \| [Winobias 1_2](https://arxiv.org/abs/1804.06876) \| \| 56.12 \| 59.09 \| \| [Winobias 2_2](https://arxiv.org/abs/1804.06876) \| \| 91.10 \| 92.23 \| \| [Toxigen](https://arxiv.org/abs/2203.09509) \| \| 29.77 \| 39.59 \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit](http://ai.google.dev/gemma/responsible). * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques. ### Benefits At the time of release, this family of models provides high-performance open large language model implementations designed from the ground up for Responsible AI development compared to similarly sized models. Using the benchmark evaluation metrics described in this document, these models have shown to provide superior performance to other, comparably-sized open model alternatives.	null	Non_BioNLP	AWQ quantized version of gemma-7b model. --- # Gemma Model Card Model Page: [Gemma](https://ai.google.dev/gemma/docs) This model card corresponds to the 7B base version of the Gemma model. You can also visit the model card of the [2B base model](https://huggingface.co/google/gemma-2b), [7B instruct model](https://huggingface.co/google/gemma-7b-it), and [2B instruct model](https://huggingface.co/google/gemma-2b-it). Resources and Technical Documentation: * [Responsible Generative AI Toolkit](https://ai.google.dev/responsible) * [Gemma on Kaggle](https://www.kaggle.com/models/google/gemma) * [Gemma on Vertex Model Garden](https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335?version=gemma-7b-gg-hf) Terms of Use: [Terms](https://www.kaggle.com/models/google/gemma/license/consent) Authors: Google ## Model Information Summary description and brief definition of inputs and outputs. ### Description Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. They are text-to-text, decoder-only large language models, available in English, with open weights, pre-trained variants, and instruction-tuned variants. Gemma models are well-suited for a variety of text generation tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as a laptop, desktop or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. ### Usage Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase. #### Fine-tuning examples You can find fine-tuning notebooks under the [`examples/` directory](https://huggingface.co/google/gemma-7b/tree/main/examples). We provide: * A script to perform Supervised Fine-Tuning (SFT) on UltraChat dataset using [QLoRA](https://huggingface.co/papers/2305.14314) * A script to perform SFT using FSDP on TPU devices * A notebook that you can run on a free-tier Google Colab instance to perform SFT on English quotes dataset. You can also find the copy of the notebook [here](https://github.com/huggingface/notebooks/blob/main/peft/gemma_7b_english_quotes.ipynb). #### Running the model on a CPU ```python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt") outputs = model.generate(input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a GPU using different precisions * _Using `torch.float16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.float16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` _Using `torch.bfloat16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.bfloat16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` _Using 8-bit precision (int8)_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` _Using 4-bit precision_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(*input_ids) print(tokenizer.decode(outputs[0])) ``` #### Other optimizations _Flash Attention 2_ First make sure to install `flash-attn` in your environment `pip install flash-attn` ```diff model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, + attn_implementation="flash_attention_2" ).to(0) ``` ### Inputs and outputs * Input: Text string, such as a question, a prompt, or a document to be summarized. * Output: Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ## Model Data Data used for model training and how the data was processed. ### Training Dataset These models were trained on a dataset of text data that includes a wide variety of sources, totaling 6 trillion tokens. Here are the key components: * Web Documents: A diverse collection of web text ensures the model is exposed to a broad range of linguistic styles, topics, and vocabulary. Primarily English-language content. * Code: Exposing the model to code helps it to learn the syntax and patterns of programming languages, which improves its ability to generate code or understand code-related questions. * Mathematics: Training on mathematical text helps the model learn logical reasoning, symbolic representation, and to address mathematical queries. The combination of these diverse data sources is crucial for training a powerful language model that can handle a wide variety of different tasks and text formats. ### Data Preprocessing Here are the key data cleaning and filtering methods applied to the training data: * CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering was applied at multiple stages in the data preparation process to ensure the exclusion of harmful and illegal content * Sensitive Data Filtering: As part of making Gemma pre-trained models safe and reliable, automated techniques were used to filter out certain personal information and other sensitive data from training sets. * Additional methods: Filtering based on content quality and safely in line with [our policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11). ## Implementation Information Details about the model internals. ### Hardware Gemma was trained using the latest generation of [Tensor Processing Unit (TPU)](https://cloud.google.com/tpu/docs/intro-to-tpu) hardware (TPUv5e). Training large language models requires significant computational power. TPUs, designed specifically for matrix operations common in machine learning, offer several advantages in this domain: * Performance: TPUs are specifically designed to handle the massive computations involved in training LLMs. They can speed up training considerably compared to CPUs. * Memory: TPUs often come with large amounts of high-bandwidth memory, allowing for the handling of large models and batch sizes during training. This can lead to better model quality. * Scalability: TPU Pods (large clusters of TPUs) provide a scalable solution for handling the growing complexity of large foundation models. You can distribute training across multiple TPU devices for faster and more efficient processing. * Cost-effectiveness: In many scenarios, TPUs can provide a more cost-effective solution for training large models compared to CPU-based infrastructure, especially when considering the time and resources saved due to faster training. * These advantages are aligned with [Google's commitments to operate sustainably](https://sustainability.google/operating-sustainably/). ### Software Training was done using [JAX](https://github.com/google/jax) and [ML Pathways](https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture). JAX allows researchers to take advantage of the latest generation of hardware, including TPUs, for faster and more efficient training of large models. ML Pathways is Google's latest effort to build artificially intelligent systems capable of generalizing across multiple tasks. This is specially suitable for [foundation models](https://ai.google/discover/foundation-models/), including large language models like these ones. Together, JAX and ML Pathways are used as described in the [paper about the Gemini family of models](https://arxiv.org/abs/2312.11805); "the 'single controller' programming model of Jax and Pathways allows a single Python process to orchestrate the entire training run, dramatically simplifying the development workflow." ## Evaluation Model evaluation metrics and results. ### Benchmark Results These models were evaluated against a large collection of different datasets and metrics to cover different aspects of text generation: \| Benchmark \| Metric \| 2B Params \| 7B Params \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| \| [MMLU](https://arxiv.org/abs/2009.03300) \| 5-shot, top-1 \| 42.3 \| 64.3 \| \| [HellaSwag](https://arxiv.org/abs/1905.07830) \| 0-shot \|71.4 \| 81.2 \| \| [PIQA](https://arxiv.org/abs/1911.11641) \| 0-shot \| 77.3 \| 81.2 \| \| [SocialIQA](https://arxiv.org/abs/1904.09728) \| 0-shot \| 59.7 \| 51.8 \| \| [BooIQ](https://arxiv.org/abs/1905.10044) \| 0-shot \| 69.4 \| 83.2 \| \| [WinoGrande](https://arxiv.org/abs/1907.10641) \| partial score \| 65.4 \| 72.3 \| \| [CommonsenseQA](https://arxiv.org/abs/1811.00937) \| 7-shot \| 65.3 \| 71.3 \| \| [OpenBookQA](https://arxiv.org/abs/1809.02789) \| \| 47.8 \| 52.8 \| \| [ARC-e](https://arxiv.org/abs/1911.01547) \| \| 73.2 \| 81.5 \| \| [ARC-c](https://arxiv.org/abs/1911.01547) \| \| 42.1 \| 53.2 \| \| [TriviaQA](https://arxiv.org/abs/1705.03551) \| 5-shot \| 53.2 \| 63.4 \| \| [Natural Questions](https://github.com/google-research-datasets/natural-questions) \| 5-shot \| - \| 23 \| \| [HumanEval](https://arxiv.org/abs/2107.03374) \| pass@1 \| 22.0 \| 32.3 \| \| [MBPP](https://arxiv.org/abs/2108.07732) \| 3-shot \| 29.2 \| 44.4 \| \| [GSM8K](https://arxiv.org/abs/2110.14168) \| maj@1 \| 17.7 \| 46.4 \| \| [MATH](https://arxiv.org/abs/2108.07732) \| 4-shot \| 11.8 \| 24.3 \| \| [AGIEval](https://arxiv.org/abs/2304.06364) \| \| 24.2 \| 41.7 \| \| [BIG-Bench](https://arxiv.org/abs/2206.04615) \| \| 35.2 \| 55.1 \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| \| Average \| \| 54.0 \| 56.4 \| ## Ethics and Safety Ethics and safety evaluation approach and results. ### Evaluation Approach Our evaluation methods include structured evaluations and internal red-teaming testing of relevant content policies. Red-teaming was conducted by a number of different teams, each with different goals and human evaluation metrics. These models were evaluated against a number of different categories relevant to ethics and safety, including: * Text-to-Text Content Safety: Human evaluation on prompts covering safety policies including child sexual abuse and exploitation, harassment, violence and gore, and hate speech. * Text-to-Text Representational Harms: Benchmark against relevant academic datasets such as [WinoBias](https://arxiv.org/abs/1804.06876) and [BBQ Dataset](https://arxiv.org/abs/2110.08193v2). * Memorization: Automated evaluation of memorization of training data, including the risk of personally identifiable information exposure. * Large-scale harm: Tests for "dangerous capabilities," such as chemical, biological, radiological, and nuclear (CBRN) risks. ### Evaluation Results The results of ethics and safety evaluations are within acceptable thresholds for meeting [internal policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11) for categories such as child safety, content safety, representational harms, memorization, large-scale harms. On top of robust internal evaluations, the results of well known safety benchmarks like BBQ, BOLD, Winogender, Winobias, RealToxicity, and TruthfulQA are shown here. \| Benchmark \| Metric \| 2B Params \| 7B Params \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| \| [RealToxicity](https://arxiv.org/abs/2009.11462) \| average \| 6.86 \| 7.90 \| \| [BOLD](https://arxiv.org/abs/2101.11718) \| \| 45.57 \| 49.08 \| \| [CrowS-Pairs](https://aclanthology.org/2020.emnlp-main.154/) \| top-1 \| 45.82 \| 51.33 \| \| [BBQ Ambig](https://arxiv.org/abs/2110.08193v2) \| 1-shot, top-1 \| 62.58 \| 92.54 \| \| [BBQ Disambig](https://arxiv.org/abs/2110.08193v2) \| top-1 \| 54.62 \| 71.99 \| \| [Winogender](https://arxiv.org/abs/1804.09301) \| top-1 \| 51.25 \| 54.17 \| \| [TruthfulQA](https://arxiv.org/abs/2109.07958) \| \| 44.84 \| 31.81 \| \| [Winobias 1_2](https://arxiv.org/abs/1804.06876) \| \| 56.12 \| 59.09 \| \| [Winobias 2_2](https://arxiv.org/abs/1804.06876) \| \| 91.10 \| 92.23 \| \| [Toxigen](https://arxiv.org/abs/2203.09509) \| \| 29.77 \| 39.59 \| \| ------------------------------ \| ------------- \| ----------- \| --------- \| ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit](http://ai.google.dev/gemma/responsible). * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques. ### Benefits At the time of release, this family of models provides high-performance open large language model implementations designed from the ground up for Responsible AI development compared to similarly sized models. Using the benchmark evaluation metrics described in this document, these models have shown to provide superior performance to other, comparably-sized open model alternatives.	{"library_name": "transformers", "license": "other", "license_name": "gemma-terms-of-use", "license_link": "https://ai.google.dev/gemma/terms", "tags": [], "extra_gated_heading": "Access Gemma on Hugging Face", "extra_gated_prompt": "To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged-in to Hugging Face and click below. Requests are processed immediately.", "extra_gated_button_content": "Acknowledge license"}	task	[ "QUESTION_ANSWERING", "SUMMARIZATION" ]	44,544
junzai/demotest	junzai	text-classification	[ "transformers", "pytorch", "bert", "text-classification", "generated_from_trainer", "en", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:05Z	2022-02-23T07:51:36+00:00	120	0	--- datasets: - glue language: - en license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: bert_finetuning_test results: - task: type: text-classification name: Text Classification dataset: name: GLUE MRPC type: glue args: mrpc metrics: - type: accuracy value: 0.8284313725490197 name: Accuracy - type: f1 value: 0.8817567567567567 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert_finetuning_test This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.4023 - Accuracy: 0.8284 - F1: 0.8818 - Combined Score: 0.8551 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1.0 ### Training results ### Framework versions - Transformers 4.16.0.dev0 - Pytorch 1.10.1+cu102 - Datasets 1.17.0 - Tokenizers 0.11.0	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert_finetuning_test This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.4023 - Accuracy: 0.8284 - F1: 0.8818 - Combined Score: 0.8551 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1.0 ### Training results ### Framework versions - Transformers 4.16.0.dev0 - Pytorch 1.10.1+cu102 - Datasets 1.17.0 - Tokenizers 0.11.0	{"datasets": ["glue"], "language": ["en"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert_finetuning_test", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "GLUE MRPC", "type": "glue", "args": "mrpc"}, "metrics": [{"type": "accuracy", "value": 0.8284313725490197, "name": "Accuracy"}, {"type": "f1", "value": 0.8817567567567567, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,545
skrh/bert_finetuning-sentiment-model-all-samples-normal	skrh	text-classification	[ "transformers", "tensorboard", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "dataset:imdb", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-11-23T14:26:40Z	2023-11-23T16:31:33+00:00	5	0	--- base_model: distilbert-base-uncased datasets: - imdb license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: bert_finetuning-sentiment-model-all-samples-normal results: - task: type: text-classification name: Text Classification dataset: name: imdb type: imdb config: plain_text split: test args: plain_text metrics: - type: accuracy value: 0.93212 name: Accuracy - type: f1 value: 0.9323985181054057 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert_finetuning-sentiment-model-all-samples-normal This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2325 - Accuracy: 0.9321 - F1: 0.9324 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.35.2 - Pytorch 2.1.0+cu118 - Datasets 2.15.0 - Tokenizers 0.15.0	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert_finetuning-sentiment-model-all-samples-normal This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2325 - Accuracy: 0.9321 - F1: 0.9324 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.35.2 - Pytorch 2.1.0+cu118 - Datasets 2.15.0 - Tokenizers 0.15.0	{"base_model": "distilbert-base-uncased", "datasets": ["imdb"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert_finetuning-sentiment-model-all-samples-normal", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "imdb", "type": "imdb", "config": "plain_text", "split": "test", "args": "plain_text"}, "metrics": [{"type": "accuracy", "value": 0.93212, "name": "Accuracy"}, {"type": "f1", "value": 0.9323985181054057, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,546
YakovElm/Hyperledger15SetFitModel_balance_ratio_Half	YakovElm	text-classification	[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	2023-06-01T14:27:59Z	2023-06-01T14:28:34+00:00	10	0	--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- # YakovElm/Hyperledger15SetFitModel_balance_ratio_Half This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("YakovElm/Hyperledger15SetFitModel_balance_ratio_Half") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	null	Non_BioNLP	# YakovElm/Hyperledger15SetFitModel_balance_ratio_Half This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("YakovElm/Hyperledger15SetFitModel_balance_ratio_Half") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}	task	[ "TEXT_CLASSIFICATION" ]	44,547
navjordj/snl-large-summarization	navjordj	text2text-generation	[ "transformers", "pytorch", "t5", "text2text-generation", "generated_from_trainer", "dataset:navjordj/SNL_summarization", "license:apache-2.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2023-02-07T13:41:51Z	2023-02-08T12:02:00+00:00	19	1	--- datasets: - navjordj/SNL_summarization license: apache-2.0 tags: - generated_from_trainer inference: parameters: max_length: 160 model-index: - name: snl-large-summarization results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # snl-large-summarization This model is a fine-tuned version of [north/t5_large_NCC_lm](https://huggingface.co/north/t5_large_NCC_lm) on the navjordj/SNL_summarization dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 32 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10.0 ### Training results ### Framework versions - Transformers 4.27.0.dev0 - Pytorch 1.13.1 - Datasets 2.9.0 - Tokenizers 0.13.2	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # snl-large-summarization This model is a fine-tuned version of [north/t5_large_NCC_lm](https://huggingface.co/north/t5_large_NCC_lm) on the navjordj/SNL_summarization dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 32 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10.0 ### Training results ### Framework versions - Transformers 4.27.0.dev0 - Pytorch 1.13.1 - Datasets 2.9.0 - Tokenizers 0.13.2	{"datasets": ["navjordj/SNL_summarization"], "license": "apache-2.0", "tags": ["generated_from_trainer"], "inference": {"parameters": {"max_length": 160}}, "model-index": [{"name": "snl-large-summarization", "results": []}]}	task	[ "SUMMARIZATION" ]	44,548
Shivam098/opt-translation	Shivam098	text2text-generation	[ "transformers", "pytorch", "safetensors", "mbart", "text2text-generation", "generated_from_trainer", "dataset:opus100", "base_model:Shivam098/opt-translation", "base_model:quantized:Shivam098/opt-translation", "autotrain_compatible", "endpoints_compatible", "4-bit", "gptq", "region:us" ]	2023-08-29T07:52:03Z	2023-09-05T06:46:59+00:00	53	0	--- base_model: Shivam098/opt-translation datasets: - opus100 tags: - generated_from_trainer model-index: - name: opt-Translation results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # opt-Translation This model is a fine-tuned version of [Shivam098/opt-translation](https://huggingface.co/Shivam098/opt-translation) on the opus100 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 300 ### Training results ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # opt-Translation This model is a fine-tuned version of [Shivam098/opt-translation](https://huggingface.co/Shivam098/opt-translation) on the opus100 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 300 ### Training results ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3	{"base_model": "Shivam098/opt-translation", "datasets": ["opus100"], "tags": ["generated_from_trainer"], "model-index": [{"name": "opt-Translation", "results": []}]}	task	[ "TRANSLATION" ]	44,549
RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf	RichardErkhov	null	[ "gguf", "endpoints_compatible", "region:us" ]	2024-05-31T01:19:56Z	2024-05-31T05:35:50+00:00	4	0	--- {} --- Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) Gugugo-koen-7B-V1.1 - GGUF - Model creator: https://huggingface.co/squarelike/ - Original model: https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1/ \| Name \| Quant method \| Size \| \| ---- \| ---- \| ---- \| \| [Gugugo-koen-7B-V1.1.Q2_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q2_K.gguf) \| Q2_K \| 2.42GB \| \| [Gugugo-koen-7B-V1.1.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_XS.gguf) \| IQ3_XS \| 2.67GB \| \| [Gugugo-koen-7B-V1.1.IQ3_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_S.gguf) \| IQ3_S \| 2.81GB \| \| [Gugugo-koen-7B-V1.1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_S.gguf) \| Q3_K_S \| 2.81GB \| \| [Gugugo-koen-7B-V1.1.IQ3_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_M.gguf) \| IQ3_M \| 2.97GB \| \| [Gugugo-koen-7B-V1.1.Q3_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K.gguf) \| Q3_K \| 3.14GB \| \| [Gugugo-koen-7B-V1.1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_M.gguf) \| Q3_K_M \| 3.14GB \| \| [Gugugo-koen-7B-V1.1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_L.gguf) \| Q3_K_L \| 3.42GB \| \| [Gugugo-koen-7B-V1.1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_XS.gguf) \| IQ4_XS \| 3.47GB \| \| [Gugugo-koen-7B-V1.1.Q4_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_0.gguf) \| Q4_0 \| 3.64GB \| \| [Gugugo-koen-7B-V1.1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_NL.gguf) \| IQ4_NL \| 3.66GB \| \| [Gugugo-koen-7B-V1.1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_S.gguf) \| Q4_K_S \| 3.67GB \| \| [Gugugo-koen-7B-V1.1.Q4_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K.gguf) \| Q4_K \| 3.88GB \| \| [Gugugo-koen-7B-V1.1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_M.gguf) \| Q4_K_M \| 3.88GB \| \| [Gugugo-koen-7B-V1.1.Q4_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_1.gguf) \| Q4_1 \| 4.03GB \| \| [Gugugo-koen-7B-V1.1.Q5_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_0.gguf) \| Q5_0 \| 4.42GB \| \| [Gugugo-koen-7B-V1.1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_S.gguf) \| Q5_K_S \| 4.42GB \| \| [Gugugo-koen-7B-V1.1.Q5_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K.gguf) \| Q5_K \| 2.16GB \| \| [Gugugo-koen-7B-V1.1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_M.gguf) \| Q5_K_M \| 0.65GB \| \| [Gugugo-koen-7B-V1.1.Q5_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_1.gguf) \| Q5_1 \| 0.62GB \| \| [Gugugo-koen-7B-V1.1.Q6_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q6_K.gguf) \| Q6_K \| 0.56GB \| \| [Gugugo-koen-7B-V1.1.Q8_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q8_0.gguf) \| Q8_0 \| 0.63GB \| Original model description: --- license: apache-2.0 datasets: - squarelike/sharegpt_deepl_ko_translation language: - en - ko pipeline_tag: translation --- # Gugugo-koen-7B-V1.1 Detail repo: [https://github.com/jwj7140/Gugugo](https://github.com/jwj7140/Gugugo) ![Gugugo](./logo.png) Base Model: [Llama-2-ko-7b](https://huggingface.co/beomi/llama-2-ko-7b) Training Dataset: [sharegpt_deepl_ko_translation](https://huggingface.co/datasets/squarelike/sharegpt_deepl_ko_translation). I trained with 1x A6000 GPUs for 90 hours. ## Prompt Template KO->EN ``` ### 한국어: {sentence}</끝> ### 영어: ``` EN->KO ``` ### 영어: {sentence}</끝> ### 한국어: ``` There are GPTQ, AWQ, and GGUF support. [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF) ## Implementation Code ```python from transformers import AutoModelForCausalLM, AutoTokenizer, StoppingCriteria, StoppingCriteriaList import torch repo = "squarelike/Gugugo-koen-7B-V1.1" model = AutoModelForCausalLM.from_pretrained( repo, load_in_4bit=True device_map='auto' ) tokenizer = AutoTokenizer.from_pretrained(repo) class StoppingCriteriaSub(StoppingCriteria): def __init__(self, stops = [], encounters=1): super().__init__() self.stops = [stop for stop in stops] def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor): for stop in self.stops: if torch.all((stop == input_ids[0][-len(stop):])).item(): return True return False stop_words_ids = torch.tensor([[829, 45107, 29958], [1533, 45107, 29958], [829, 45107, 29958], [21106, 45107, 29958]]).to("cuda") stopping_criteria = StoppingCriteriaList([StoppingCriteriaSub(stops=stop_words_ids)]) def gen(lan="en", x=""): if (lan == "ko"): prompt = f"### 한국어: {x}</끝>\n### 영어:" else: prompt = f"### 영어: {x}</끝>\n### 한국어:" gened = model.generate( **tokenizer( prompt, return_tensors='pt', return_token_type_ids=False ).to("cuda"), max_new_tokens=2000, temperature=0.3, # no_repeat_ngram_size=5, num_beams=5, stopping_criteria=stopping_criteria ) return tokenizer.decode(gened[0][1:]).replace(prompt+" ", "").replace("</끝>", "") print(gen(lan="en", x="Hello, world!")) ```	null	Non_BioNLP	Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) Gugugo-koen-7B-V1.1 - GGUF - Model creator: https://huggingface.co/squarelike/ - Original model: https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1/ \| Name \| Quant method \| Size \| \| ---- \| ---- \| ---- \| \| [Gugugo-koen-7B-V1.1.Q2_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q2_K.gguf) \| Q2_K \| 2.42GB \| \| [Gugugo-koen-7B-V1.1.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_XS.gguf) \| IQ3_XS \| 2.67GB \| \| [Gugugo-koen-7B-V1.1.IQ3_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_S.gguf) \| IQ3_S \| 2.81GB \| \| [Gugugo-koen-7B-V1.1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_S.gguf) \| Q3_K_S \| 2.81GB \| \| [Gugugo-koen-7B-V1.1.IQ3_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_M.gguf) \| IQ3_M \| 2.97GB \| \| [Gugugo-koen-7B-V1.1.Q3_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K.gguf) \| Q3_K \| 3.14GB \| \| [Gugugo-koen-7B-V1.1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_M.gguf) \| Q3_K_M \| 3.14GB \| \| [Gugugo-koen-7B-V1.1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_L.gguf) \| Q3_K_L \| 3.42GB \| \| [Gugugo-koen-7B-V1.1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_XS.gguf) \| IQ4_XS \| 3.47GB \| \| [Gugugo-koen-7B-V1.1.Q4_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_0.gguf) \| Q4_0 \| 3.64GB \| \| [Gugugo-koen-7B-V1.1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_NL.gguf) \| IQ4_NL \| 3.66GB \| \| [Gugugo-koen-7B-V1.1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_S.gguf) \| Q4_K_S \| 3.67GB \| \| [Gugugo-koen-7B-V1.1.Q4_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K.gguf) \| Q4_K \| 3.88GB \| \| [Gugugo-koen-7B-V1.1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_M.gguf) \| Q4_K_M \| 3.88GB \| \| [Gugugo-koen-7B-V1.1.Q4_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_1.gguf) \| Q4_1 \| 4.03GB \| \| [Gugugo-koen-7B-V1.1.Q5_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_0.gguf) \| Q5_0 \| 4.42GB \| \| [Gugugo-koen-7B-V1.1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_S.gguf) \| Q5_K_S \| 4.42GB \| \| [Gugugo-koen-7B-V1.1.Q5_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K.gguf) \| Q5_K \| 2.16GB \| \| [Gugugo-koen-7B-V1.1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_M.gguf) \| Q5_K_M \| 0.65GB \| \| [Gugugo-koen-7B-V1.1.Q5_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_1.gguf) \| Q5_1 \| 0.62GB \| \| [Gugugo-koen-7B-V1.1.Q6_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q6_K.gguf) \| Q6_K \| 0.56GB \| \| [Gugugo-koen-7B-V1.1.Q8_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q8_0.gguf) \| Q8_0 \| 0.63GB \| Original model description: --- license: apache-2.0 datasets: - squarelike/sharegpt_deepl_ko_translation language: - en - ko pipeline_tag: translation --- # Gugugo-koen-7B-V1.1 Detail repo: [https://github.com/jwj7140/Gugugo](https://github.com/jwj7140/Gugugo) ![Gugugo](./logo.png) Base Model: [Llama-2-ko-7b](https://huggingface.co/beomi/llama-2-ko-7b) Training Dataset: [sharegpt_deepl_ko_translation](https://huggingface.co/datasets/squarelike/sharegpt_deepl_ko_translation). I trained with 1x A6000 GPUs for 90 hours. ## Prompt Template KO->EN ``` ### 한국어: {sentence}</끝> ### 영어: ``` EN->KO ``` ### 영어: {sentence}</끝> ### 한국어: ``` There are GPTQ, AWQ, and GGUF support. [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF) ## Implementation Code ```python from transformers import AutoModelForCausalLM, AutoTokenizer, StoppingCriteria, StoppingCriteriaList import torch repo = "squarelike/Gugugo-koen-7B-V1.1" model = AutoModelForCausalLM.from_pretrained( repo, load_in_4bit=True device_map='auto' ) tokenizer = AutoTokenizer.from_pretrained(repo) class StoppingCriteriaSub(StoppingCriteria): def __init__(self, stops = [], encounters=1): super().__init__() self.stops = [stop for stop in stops] def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor): for stop in self.stops: if torch.all((stop == input_ids[0][-len(stop):])).item(): return True return False stop_words_ids = torch.tensor([[829, 45107, 29958], [1533, 45107, 29958], [829, 45107, 29958], [21106, 45107, 29958]]).to("cuda") stopping_criteria = StoppingCriteriaList([StoppingCriteriaSub(stops=stop_words_ids)]) def gen(lan="en", x=""): if (lan == "ko"): prompt = f"### 한국어: {x}</끝>\n### 영어:" else: prompt = f"### 영어: {x}</끝>\n### 한국어:" gened = model.generate( **tokenizer( prompt, return_tensors='pt', return_token_type_ids=False ).to("cuda"), max_new_tokens=2000, temperature=0.3, # no_repeat_ngram_size=5, num_beams=5, stopping_criteria=stopping_criteria ) return tokenizer.decode(gened[0][1:]).replace(prompt+" ", "").replace("</끝>", "") print(gen(lan="en", x="Hello, world!")) ```	{}	task	[ "TRANSLATION" ]	44,550
pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1-sts	pritamdeka	sentence-similarity	[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:5749", "loss:CosineSimilarityLoss", "arxiv:1908.10084", "base_model:pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "base_model:finetune:pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	2024-07-17T19:02:37Z	2024-07-17T19:03:24+00:00	10	0	--- base_model: pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 datasets: [] language: [] library_name: sentence-transformers metrics: - pearson_cosine - spearman_cosine - pearson_manhattan - spearman_manhattan - pearson_euclidean - spearman_euclidean - pearson_dot - spearman_dot - pearson_max - spearman_max pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:5749 - loss:CosineSimilarityLoss widget: - source_sentence: আমি "... comoving মহাজাগতিক বিশ্ৰাম ফ্ৰেমৰ তুলনাত ... সিংহ নক্ষত্ৰমণ্ডলৰ ফালে কিছু 371 কিলোমিটাৰ প্ৰতি ছেকেণ্ডত" আগবাঢ়িছো. sentences: - বাস্কেটবল খেলুৱৈগৰাকীয়ে নিজৰ দলৰ হৈ পইণ্ট লাভ কৰিবলৈ ওলাইছে। - আন কোনো বস্তুৰ লগত আপেক্ষিক নহোৱা কোনো ‘ষ্টিল’ নাই। - এজনী ছোৱালীয়ে বতাহ বাদ্যযন্ত্ৰ বজায়। - source_sentence: চাৰিটা ল’ৰা-ছোৱালীয়ে ভঁৰালৰ জীৱ-জন্তুবোৰলৈ চাই আছে। sentences: - ডাইনিং টেবুল এখনৰ চাৰিওফালে বৃদ্ধৰ দল এটাই পোজ দিছে। - বিকিনি পিন্ধা চাৰিগৰাকী মহিলাই বিলত ভলীবল খেলি আছে। - ল’ৰা-ছোৱালীয়ে ভেড়া চাই। - source_sentence: ডালত বহি থকা দুটা টান ঈগল। sentences: - জাতৰ জেব্ৰা ডানিঅ’ অত্যন্ত কঠোৰ মাছ, ইহঁতক হত্যা কৰাটো প্ৰায় কঠিন। - এটা ডালত দুটা ঈগল বহি আছে। - নূন্যতম মজুৰিৰ আইনসমূহে কম দক্ষ, কম উৎপাদনশীল লোকক আটাইতকৈ বেছি আঘাত দিয়ে। - source_sentence: '"মই আচলতে যি বিচাৰিছো সেয়া হৈছে মুছলমান জনসংখ্যাৰ এটা অনুমান..." @ThanosK আৰু @T.E.D., এটা সামগ্ৰিক, সাধাৰণ জনসংখ্যাৰ অনুমান f.e.' sentences: - এগৰাকী মহিলাই সেউজীয়া পিঁয়াজ কাটি আছে। - তলত দিয়া কথাখিনি মোৰ কুকুৰ কাণৰ দৰে কপিৰ পৰা লোৱা হৈছে নিউ পেংগুইন এটলাছ অৱ মেডিভেল হিষ্ট্ৰীৰ। - আমাৰ দৰে সৌৰজগতৰ কোনো তাৰকাৰাজ্যৰ বাহিৰত থকাটো সম্ভৱ হ’ব পাৰে। - source_sentence: ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে। sentences: - গছৰ শাৰী এটাৰ সন্মুখত পথাৰত ভেড়া চৰিছে। - এজন মানুহে গীটাৰ বজাই আছে। - ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে। model-index: - name: SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 results: - task: type: semantic-similarity name: Semantic Similarity dataset: name: pritamdeka/stsb assamese translated dev type: pritamdeka/stsb-assamese-translated-dev metrics: - type: pearson_cosine value: 0.8525258323169252 name: Pearson Cosine - type: spearman_cosine value: 0.8506593647943235 name: Spearman Cosine - type: pearson_manhattan value: 0.8334889460288037 name: Pearson Manhattan - type: spearman_manhattan value: 0.843042040822402 name: Spearman Manhattan - type: pearson_euclidean value: 0.8351723933495433 name: Pearson Euclidean - type: spearman_euclidean value: 0.8450734552112781 name: Spearman Euclidean - type: pearson_dot value: 0.8273071926204811 name: Pearson Dot - type: spearman_dot value: 0.8277520425148079 name: Spearman Dot - type: pearson_max value: 0.8525258323169252 name: Pearson Max - type: spearman_max value: 0.8506593647943235 name: Spearman Max - task: type: semantic-similarity name: Semantic Similarity dataset: name: pritamdeka/stsb assamese translated test type: pritamdeka/stsb-assamese-translated-test metrics: - type: pearson_cosine value: 0.8138083526567048 name: Pearson Cosine - type: spearman_cosine value: 0.8119367763029309 name: Spearman Cosine - type: pearson_manhattan value: 0.8044112753419641 name: Pearson Manhattan - type: spearman_manhattan value: 0.8073243490029997 name: Spearman Manhattan - type: pearson_euclidean value: 0.805728285628756 name: Pearson Euclidean - type: spearman_euclidean value: 0.8086070843216111 name: Spearman Euclidean - type: pearson_dot value: 0.7754575809083841 name: Pearson Dot - type: spearman_dot value: 0.7720173359758135 name: Spearman Dot - type: pearson_max value: 0.8138083526567048 name: Pearson Max - type: spearman_max value: 0.8119367763029309 name: Spearman Max --- # SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1) <!-- at revision ec876d6ec1d2030ad233470e574f1d3d3fe56c74 --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 768 tokens - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1-sts") # Run inference sentences = [ 'ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে।', 'ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে।', 'এজন মানুহে গীটাৰ বজাই আছে।', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-dev` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| pearson_cosine \| 0.8525 \| \| spearman_cosine \| 0.8507 \| \| pearson_manhattan \| 0.8335 \| \| spearman_manhattan \| 0.843 \| \| pearson_euclidean \| 0.8352 \| \| spearman_euclidean \| 0.8451 \| \| pearson_dot \| 0.8273 \| \| spearman_dot \| 0.8278 \| \| pearson_max \| 0.8525 \| \| spearman_max \| 0.8507 \| #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-test` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| pearson_cosine \| 0.8138 \| \| spearman_cosine \| 0.8119 \| \| pearson_manhattan \| 0.8044 \| \| spearman_manhattan \| 0.8073 \| \| pearson_euclidean \| 0.8057 \| \| spearman_euclidean \| 0.8086 \| \| pearson_dot \| 0.7755 \| \| spearman_dot \| 0.772 \| \| pearson_max \| 0.8138 \| \| spearman_max \| 0.8119 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `num_train_epochs`: 10 - `warmup_ratio`: 0.1 - `fp16`: True - `load_best_model_at_end`: True #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: True - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs \| Epoch \| Step \| Training Loss \| loss \| pritamdeka/stsb-assamese-translated-dev_spearman_cosine \| pritamdeka/stsb-assamese-translated-test_spearman_cosine \| \|:----------:\|:-------:\|:-------------:\|:----------:\|:-------------------------------------------------------:\|:--------------------------------------------------------:\| \| 1.1111 \| 100 \| 0.0331 \| 0.0259 \| 0.8482 \| - \| \| 2.2222 \| 200 \| 0.0176 \| 0.0253 \| 0.8515 \| - \| \| 3.3333 \| 300 \| 0.011 \| 0.0253 \| 0.8513 \| - \| \| 4.4444 \| 400 \| 0.0066 \| 0.0259 \| 0.8492 \| - \| \| 5.5556 \| 500 \| 0.0048 \| 0.0255 \| 0.8511 \| - \| \| 6.6667 \| 600 \| 0.0037 \| 0.0256 \| 0.8508 \| - \| \| 7.7778 \| 700 \| 0.0033 \| 0.0254 \| 0.8515 \| - \| \| 8.8889 \| 800 \| 0.0029 \| 0.0255 \| 0.8512 \| - \| \| 10.0 \| 900 \| 0.0027 \| 0.0257 \| 0.8507 \| 0.8119 \| * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.12 - Sentence Transformers: 3.0.1 - Transformers: 4.42.4 - PyTorch: 2.3.1+cu121 - Accelerate: 0.32.1 - Datasets: 2.20.0 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1) <!-- at revision ec876d6ec1d2030ad233470e574f1d3d3fe56c74 --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 768 tokens - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1-sts") # Run inference sentences = [ 'ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে।', 'ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে।', 'এজন মানুহে গীটাৰ বজাই আছে।', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-dev` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| pearson_cosine \| 0.8525 \| \| spearman_cosine \| 0.8507 \| \| pearson_manhattan \| 0.8335 \| \| spearman_manhattan \| 0.843 \| \| pearson_euclidean \| 0.8352 \| \| spearman_euclidean \| 0.8451 \| \| pearson_dot \| 0.8273 \| \| spearman_dot \| 0.8278 \| \| pearson_max \| 0.8525 \| \| spearman_max \| 0.8507 \| #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-test` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| pearson_cosine \| 0.8138 \| \| spearman_cosine \| 0.8119 \| \| pearson_manhattan \| 0.8044 \| \| spearman_manhattan \| 0.8073 \| \| pearson_euclidean \| 0.8057 \| \| spearman_euclidean \| 0.8086 \| \| pearson_dot \| 0.7755 \| \| spearman_dot \| 0.772 \| \| pearson_max \| 0.8138 \| \| spearman_max \| 0.8119 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `num_train_epochs`: 10 - `warmup_ratio`: 0.1 - `fp16`: True - `load_best_model_at_end`: True #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: True - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs \| Epoch \| Step \| Training Loss \| loss \| pritamdeka/stsb-assamese-translated-dev_spearman_cosine \| pritamdeka/stsb-assamese-translated-test_spearman_cosine \| \|:----------:\|:-------:\|:-------------:\|:----------:\|:-------------------------------------------------------:\|:--------------------------------------------------------:\| \| 1.1111 \| 100 \| 0.0331 \| 0.0259 \| 0.8482 \| - \| \| 2.2222 \| 200 \| 0.0176 \| 0.0253 \| 0.8515 \| - \| \| 3.3333 \| 300 \| 0.011 \| 0.0253 \| 0.8513 \| - \| \| 4.4444 \| 400 \| 0.0066 \| 0.0259 \| 0.8492 \| - \| \| 5.5556 \| 500 \| 0.0048 \| 0.0255 \| 0.8511 \| - \| \| 6.6667 \| 600 \| 0.0037 \| 0.0256 \| 0.8508 \| - \| \| 7.7778 \| 700 \| 0.0033 \| 0.0254 \| 0.8515 \| - \| \| 8.8889 \| 800 \| 0.0029 \| 0.0255 \| 0.8512 \| - \| \| 10.0 \| 900 \| 0.0027 \| 0.0257 \| 0.8507 \| 0.8119 \| * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.12 - Sentence Transformers: 3.0.1 - Transformers: 4.42.4 - PyTorch: 2.3.1+cu121 - Accelerate: 0.32.1 - Datasets: 2.20.0 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "datasets": [], "language": [], "library_name": "sentence-transformers", "metrics": ["pearson_cosine", "spearman_cosine", "pearson_manhattan", "spearman_manhattan", "pearson_euclidean", "spearman_euclidean", "pearson_dot", "spearman_dot", "pearson_max", "spearman_max"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:5749", "loss:CosineSimilarityLoss"], "widget": [{"source_sentence": "আমি \"... comoving মহাজাগতিক বিশ্ৰাম ফ্ৰেমৰ তুলনাত ... সিংহ নক্ষত্ৰমণ্ডলৰ ফালে কিছু 371 কিলোমিটাৰ প্ৰতি ছেকেণ্ডত\" আগবাঢ়িছো.", "sentences": ["বাস্কেটবল খেলুৱৈগৰাকীয়ে নিজৰ দলৰ হৈ পইণ্ট লাভ কৰিবলৈ ওলাইছে।", "আন কোনো বস্তুৰ লগত আপেক্ষিক নহোৱা কোনো ‘ষ্টিল’ নাই।", "এজনী ছোৱালীয়ে বতাহ বাদ্যযন্ত্ৰ বজায়।"]}, {"source_sentence": "চাৰিটা ল’ৰা-ছোৱালীয়ে ভঁৰালৰ জীৱ-জন্তুবোৰলৈ চাই আছে।", "sentences": ["ডাইনিং টেবুল এখনৰ চাৰিওফালে বৃদ্ধৰ দল এটাই পোজ দিছে।", "বিকিনি পিন্ধা চাৰিগৰাকী মহিলাই বিলত ভলীবল খেলি আছে।", "ল’ৰা-ছোৱালীয়ে ভেড়া চাই।"]}, {"source_sentence": "ডালত বহি থকা দুটা টান ঈগল।", "sentences": ["জাতৰ জেব্ৰা ডানিঅ’ অত্যন্ত কঠোৰ মাছ, ইহঁতক হত্যা কৰাটো প্ৰায় কঠিন।", "এটা ডালত দুটা ঈগল বহি আছে।", "নূন্যতম মজুৰিৰ আইনসমূহে কম দক্ষ, কম উৎপাদনশীল লোকক আটাইতকৈ বেছি আঘাত দিয়ে।"]}, {"source_sentence": "\"মই আচলতে যি বিচাৰিছো সেয়া হৈছে মুছলমান জনসংখ্যাৰ এটা অনুমান...\" @ThanosK আৰু @T.E.D., এটা সামগ্ৰিক, সাধাৰণ জনসংখ্যাৰ অনুমান f.e.", "sentences": ["এগৰাকী মহিলাই সেউজীয়া পিঁয়াজ কাটি আছে।", "তলত দিয়া কথাখিনি মোৰ কুকুৰ কাণৰ দৰে কপিৰ পৰা লোৱা হৈছে নিউ পেংগুইন এটলাছ অৱ মেডিভেল হিষ্ট্ৰীৰ।", "আমাৰ দৰে সৌৰজগতৰ কোনো তাৰকাৰাজ্যৰ বাহিৰত থকাটো সম্ভৱ হ’ব পাৰে।"]}, {"source_sentence": "ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে।", "sentences": ["গছৰ শাৰী এটাৰ সন্মুখত পথাৰত ভেড়া চৰিছে।", "এজন মানুহে গীটাৰ বজাই আছে।", "ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে।"]}], "model-index": [{"name": "SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "results": [{"task": {"type": "semantic-similarity", "name": "Semantic Similarity"}, "dataset": {"name": "pritamdeka/stsb assamese translated dev", "type": "pritamdeka/stsb-assamese-translated-dev"}, "metrics": [{"type": "pearson_cosine", "value": 0.8525258323169252, "name": "Pearson Cosine"}, {"type": "spearman_cosine", "value": 0.8506593647943235, "name": "Spearman Cosine"}, {"type": "pearson_manhattan", "value": 0.8334889460288037, "name": "Pearson Manhattan"}, {"type": "spearman_manhattan", "value": 0.843042040822402, "name": "Spearman Manhattan"}, {"type": "pearson_euclidean", "value": 0.8351723933495433, "name": "Pearson Euclidean"}, {"type": "spearman_euclidean", "value": 0.8450734552112781, "name": "Spearman Euclidean"}, {"type": "pearson_dot", "value": 0.8273071926204811, "name": "Pearson Dot"}, {"type": "spearman_dot", "value": 0.8277520425148079, "name": "Spearman Dot"}, {"type": "pearson_max", "value": 0.8525258323169252, "name": "Pearson Max"}, {"type": "spearman_max", "value": 0.8506593647943235, "name": "Spearman Max"}]}, {"task": {"type": "semantic-similarity", "name": "Semantic Similarity"}, "dataset": {"name": "pritamdeka/stsb assamese translated test", "type": "pritamdeka/stsb-assamese-translated-test"}, "metrics": [{"type": "pearson_cosine", "value": 0.8138083526567048, "name": "Pearson Cosine"}, {"type": "spearman_cosine", "value": 0.8119367763029309, "name": "Spearman Cosine"}, {"type": "pearson_manhattan", "value": 0.8044112753419641, "name": "Pearson Manhattan"}, {"type": "spearman_manhattan", "value": 0.8073243490029997, "name": "Spearman Manhattan"}, {"type": "pearson_euclidean", "value": 0.805728285628756, "name": "Pearson Euclidean"}, {"type": "spearman_euclidean", "value": 0.8086070843216111, "name": "Spearman Euclidean"}, {"type": "pearson_dot", "value": 0.7754575809083841, "name": "Pearson Dot"}, {"type": "spearman_dot", "value": 0.7720173359758135, "name": "Spearman Dot"}, {"type": "pearson_max", "value": 0.8138083526567048, "name": "Pearson Max"}, {"type": "spearman_max", "value": 0.8119367763029309, "name": "Spearman Max"}]}]}]}	task	[ "TEXT_CLASSIFICATION", "SEMANTIC_SIMILARITY" ]	44,551
luckyf1998/distilbert-base-uncased-finetuned-emotion	luckyf1998	text-classification	[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:emotion", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-08-03T11:31:49Z	2023-08-03T14:34:09+00:00	16	0	--- datasets: - emotion license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion args: split metrics: - type: accuracy value: 0.923 name: Accuracy - type: f1 value: 0.9227217081326218 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2148 - Accuracy: 0.923 - F1: 0.9227 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 0.8269 \| 1.0 \| 250 \| 0.3107 \| 0.909 \| 0.9072 \| \| 0.2402 \| 2.0 \| 500 \| 0.2148 \| 0.923 \| 0.9227 \| ### Framework versions - Transformers 4.16.2 - Pytorch 2.0.0 - Datasets 1.16.1 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2148 - Accuracy: 0.923 - F1: 0.9227 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 0.8269 \| 1.0 \| 250 \| 0.3107 \| 0.909 \| 0.9072 \| \| 0.2402 \| 2.0 \| 500 \| 0.2148 \| 0.923 \| 0.9227 \| ### Framework versions - Transformers 4.16.2 - Pytorch 2.0.0 - Datasets 1.16.1 - Tokenizers 0.13.3	{"datasets": ["emotion"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-emotion", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "args": "split"}, "metrics": [{"type": "accuracy", "value": 0.923, "name": "Accuracy"}, {"type": "f1", "value": 0.9227217081326218, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,552
Z3R6X/gpt4all_dpo_instruct	Z3R6X	text-generation	[ "transformers", "pytorch", "gptj", "text-generation", "en", "dataset:Dahoas/instruct-synthetic-prompt-responses", "arxiv:2305.18290", "license:cc-by-nc-4.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2024-03-10T15:27:19Z	2024-04-19T08:02:51+00:00	16	0	--- datasets: - Dahoas/instruct-synthetic-prompt-responses language: - en license: cc-by-nc-4.0 pipeline_tag: text-generation --- Question answering model finetuned from [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) with [Direct Preference Optimization](https://arxiv.org/abs/2305.18290). \ Dataset: [Dahoas/instruct-synthetic-prompt-responses](https://huggingface.co/datasets/Dahoas/instruct-synthetic-prompt-responses). The model was finetuned with the following promt: \ ``"Answer the following question in context:\n\nQuestion: " + samples["prompt"] + " Answer: "`` \ It should be benefical to use the same or a similar prompt for inference. An increase in performance compared to [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) was observed when using two-shot Chain-of-Thought prompting. \| HellaSwag \| WinoGrande \| BooLQ \| ARC-c \| \|:------:\|:------:\|:------:\|:------:\| \| 62.37% \| 63.3% \| 65.2% \| 32.76% \|	null	Non_BioNLP	Question answering model finetuned from [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) with [Direct Preference Optimization](https://arxiv.org/abs/2305.18290). \ Dataset: [Dahoas/instruct-synthetic-prompt-responses](https://huggingface.co/datasets/Dahoas/instruct-synthetic-prompt-responses). The model was finetuned with the following promt: \ ``"Answer the following question in context:\n\nQuestion: " + samples["prompt"] + " Answer: "`` \ It should be benefical to use the same or a similar prompt for inference. An increase in performance compared to [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) was observed when using two-shot Chain-of-Thought prompting. \| HellaSwag \| WinoGrande \| BooLQ \| ARC-c \| \|:------:\|:------:\|:------:\|:------:\| \| 62.37% \| 63.3% \| 65.2% \| 32.76% \|	{"datasets": ["Dahoas/instruct-synthetic-prompt-responses"], "language": ["en"], "license": "cc-by-nc-4.0", "pipeline_tag": "text-generation"}	task	[ "QUESTION_ANSWERING" ]	44,553
emozilla/LLongMA-2-13b-storysummarizer-GGML	emozilla	null	[ "dataset:emozilla/booksum-summary-analysis_llama-8192", "dataset:kmfoda/booksum", "license:apache-2.0", "region:us" ]	2023-08-07T15:28:28Z	2023-08-07T18:51:05+00:00	0	0	--- datasets: - emozilla/booksum-summary-analysis_llama-8192 - kmfoda/booksum license: apache-2.0 --- GGML version of [emozilla/LLongMA-2-13b-storysummarizer](https://huggingface.co/emozilla/LLongMA-2-13b-storysummarizer). Use `n_ctx=8192` and `rope_freq_scale=0.5`. For an excellent user interface, check out [LM Studio](https://lmstudio.ai/). # LLongMA-2-13b-storysummarizer [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl) This is a fine-tuned version of [conceptofmind/LLongMA-2-13b](https://huggingface.co/conceptofmind/LLongMA-2-13b) intended for summarization and literary analysis of fiction stories. It contains custom modeling code to use Flash Attention 2 during inference, which provides a significant speedup, especially at longer context lengths. To enable, pass `trust_remote_code=True,use_flash_attention=True` to `AutoModelForCausalLM`. ## How to Use ``` SOME_FICTION ### SUMMARY: ``` or ``` SOME_FICTION ### ANALYSIS: ``` ## Example outputs Outputs on the text of [_Waystation City_](https://www.uncannymagazine.com/article/waystation-city/) (6,904 tokens) ### Summary `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > A nineteen-seventies pair of twins approaches Micheala and Gerty at a cafe. They ask them to write down their story of their disappearance. Gerty takes notes on their story as they tell it. The twins' names are Daphne and Claude. The twins explain that they came from London, England and were leading a movement against fascism there. But in Waystation City, they were stuck in the limbo between timelines with no real way of returning home to continue their work. In fact, it was harder for them to remember their lives at home as time went on. They tell Gerty that they cannot stay and Micheala encourages them not to go down to the stairs where Waystation City's ferry drivers will take them home. Daphne writes an inscription on a wall, leaving her mark as one of the disappeared, but she and Claude relent. Together they approach the docks with Gerty and Micheala trailing behind. There, the twins discuss which hole to go through and they agree it is the one showing their hometown in the 1970s. Daphne goes first and then Claude after he slips on a wet surface. The ferry drivers see him but Gerty yells for him to return from the hole as she writes his story down with her notes. But he cannot hear her over the slip of metal slats turning into stone and wood on the docks. He jumps in anyway and disappears into the other world through the hole. Micheala leaves a cigarette and Gerty drops her ticket back home to show that she's decided to stay in Waystation City where it is possible to grow, change, and live in the spaces between. ### Analysis `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > Notes In this final chapter, there is a brief but important flashback to the twins' home in England at the beginning of the 1970s, when racism and fascism are rampant. It is interesting to note the author's choice to set the story at such a pivotal time for immigrants in Britain. The twins feel like they have to choose between being their "true" selves in their home country or staying where they are as Waystation citizens. But this was also the time when racism and fascism were so openly practiced that anti-racists were fighting against these institutions with all of their might. It is interesting how there's an implication here that it's only at the risk of their "true" identities that the twins can go back to England. In this chapter, Gerty changes her name as well, choosing the gender neutral Michael. She does not specify why she decided to change her name or what inspired her choice. The story implies, but never says outright, that she changed her name in part because of the twins' disappearance. This suggests an important question: how far must a person be willing to go in order to stay true to themselves? It is clear by this point that Gerty has changed quite a bit since coming to Waystation City and her decision to change her name reflects the extent to which she has altered her identity. She realizes that she has settled into a comfortable, new life here with new clothes, new hobbies, and friends. Ultimately, this chapter suggests that while it is important to stay true to one's past, there are also new things to discover and experience in the present. Gerty has done just this by settling in Waystation City, embracing her new life and even changing her name and identity with it. ## Training The model was trained on [emozilla/booksum-summary-analysis_llama-8192](https://huggingface.co/datasets/emozilla/booksum-summary-analysis_llama-8192), which is adapted from [kmfoda/booksum](https://huggingface.co/datasets/kmfoda/booksum). The training run was performed using [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl). The run can be viewed on [wandb](https://wandb.ai/emozilla/booksum/runs/53su11ol).	null	Non_BioNLP	GGML version of [emozilla/LLongMA-2-13b-storysummarizer](https://huggingface.co/emozilla/LLongMA-2-13b-storysummarizer). Use `n_ctx=8192` and `rope_freq_scale=0.5`. For an excellent user interface, check out [LM Studio](https://lmstudio.ai/). # LLongMA-2-13b-storysummarizer [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl) This is a fine-tuned version of [conceptofmind/LLongMA-2-13b](https://huggingface.co/conceptofmind/LLongMA-2-13b) intended for summarization and literary analysis of fiction stories. It contains custom modeling code to use Flash Attention 2 during inference, which provides a significant speedup, especially at longer context lengths. To enable, pass `trust_remote_code=True,use_flash_attention=True` to `AutoModelForCausalLM`. ## How to Use ``` SOME_FICTION ### SUMMARY: ``` or ``` SOME_FICTION ### ANALYSIS: ``` ## Example outputs Outputs on the text of [_Waystation City_](https://www.uncannymagazine.com/article/waystation-city/) (6,904 tokens) ### Summary `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > A nineteen-seventies pair of twins approaches Micheala and Gerty at a cafe. They ask them to write down their story of their disappearance. Gerty takes notes on their story as they tell it. The twins' names are Daphne and Claude. The twins explain that they came from London, England and were leading a movement against fascism there. But in Waystation City, they were stuck in the limbo between timelines with no real way of returning home to continue their work. In fact, it was harder for them to remember their lives at home as time went on. They tell Gerty that they cannot stay and Micheala encourages them not to go down to the stairs where Waystation City's ferry drivers will take them home. Daphne writes an inscription on a wall, leaving her mark as one of the disappeared, but she and Claude relent. Together they approach the docks with Gerty and Micheala trailing behind. There, the twins discuss which hole to go through and they agree it is the one showing their hometown in the 1970s. Daphne goes first and then Claude after he slips on a wet surface. The ferry drivers see him but Gerty yells for him to return from the hole as she writes his story down with her notes. But he cannot hear her over the slip of metal slats turning into stone and wood on the docks. He jumps in anyway and disappears into the other world through the hole. Micheala leaves a cigarette and Gerty drops her ticket back home to show that she's decided to stay in Waystation City where it is possible to grow, change, and live in the spaces between. ### Analysis `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > Notes In this final chapter, there is a brief but important flashback to the twins' home in England at the beginning of the 1970s, when racism and fascism are rampant. It is interesting to note the author's choice to set the story at such a pivotal time for immigrants in Britain. The twins feel like they have to choose between being their "true" selves in their home country or staying where they are as Waystation citizens. But this was also the time when racism and fascism were so openly practiced that anti-racists were fighting against these institutions with all of their might. It is interesting how there's an implication here that it's only at the risk of their "true" identities that the twins can go back to England. In this chapter, Gerty changes her name as well, choosing the gender neutral Michael. She does not specify why she decided to change her name or what inspired her choice. The story implies, but never says outright, that she changed her name in part because of the twins' disappearance. This suggests an important question: how far must a person be willing to go in order to stay true to themselves? It is clear by this point that Gerty has changed quite a bit since coming to Waystation City and her decision to change her name reflects the extent to which she has altered her identity. She realizes that she has settled into a comfortable, new life here with new clothes, new hobbies, and friends. Ultimately, this chapter suggests that while it is important to stay true to one's past, there are also new things to discover and experience in the present. Gerty has done just this by settling in Waystation City, embracing her new life and even changing her name and identity with it. ## Training The model was trained on [emozilla/booksum-summary-analysis_llama-8192](https://huggingface.co/datasets/emozilla/booksum-summary-analysis_llama-8192), which is adapted from [kmfoda/booksum](https://huggingface.co/datasets/kmfoda/booksum). The training run was performed using [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl). The run can be viewed on [wandb](https://wandb.ai/emozilla/booksum/runs/53su11ol).	{"datasets": ["emozilla/booksum-summary-analysis_llama-8192", "kmfoda/booksum"], "license": "apache-2.0"}	task	[ "SUMMARIZATION" ]	44,554
ruanchaves/bert-base-portuguese-cased-assin-entailment	ruanchaves	text-classification	[ "transformers", "pytorch", "bert", "text-classification", "pt", "dataset:assin", "autotrain_compatible", "region:us" ]	2023-03-27T18:09:12Z	2023-03-29T18:05:31+00:00	33	0	--- datasets: - assin language: pt inference: false --- # BERTimbau base for Recognizing Textual Entailment This is the [neuralmind/bert-base-portuguese-cased](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model finetuned for Recognizing Textual Entailment with the [ASSIN](https://huggingface.co/datasets/assin) dataset. This model is suitable for Portuguese. - Git Repo: [Evaluation of Portuguese Language Models](https://github.com/ruanchaves/eplm). - Demo: [Portuguese Textual Entailment](https://ruanchaves-portuguese-textual-entailment.hf.space) ### Labels: * 0 : There is no entailment between premise and hypothesis. * 1 : There is entailment between premise and hypothesis. * 2 : The premise is a paraphrase of the hypothesis. ## Full classification example ```python from transformers import AutoModelForSequenceClassification, AutoTokenizer, AutoConfig import numpy as np import torch from scipy.special import softmax model_name = "ruanchaves/bert-base-portuguese-cased-assin-entailment" s1 = "Os homens estão cuidadosamente colocando as malas no porta-malas de um carro." s2 = "Os homens estão colocando bagagens dentro do porta-malas de um carro." model = AutoModelForSequenceClassification.from_pretrained(model_name) tokenizer = AutoTokenizer.from_pretrained(model_name) config = AutoConfig.from_pretrained(model_name) model_input = tokenizer(([s1], [s2]), padding=True, return_tensors="pt") with torch.no_grad(): output = model(*model_input) scores = output[0][0].detach().numpy() scores = softmax(scores) ranking = np.argsort(scores) ranking = ranking[::-1] for i in range(scores.shape[0]): l = config.id2label[ranking[i]] s = scores[ranking[i]] print(f"{i+1}) Label: {l} Score: {np.round(float(s), 4)}") ``` ## Citation Our research is ongoing, and we are currently working on describing our experiments in a paper, which will be published soon. In the meanwhile, if you would like to cite our work or models before the publication of the paper, please cite our [GitHub repository](https://github.com/ruanchaves/eplm): ``` @software{Chaves_Rodrigues_eplm_2023, author = {Chaves Rodrigues, Ruan and Tanti, Marc and Agerri, Rodrigo}, doi = {10.5281/zenodo.7781848}, month = {3}, title = {{Evaluation of Portuguese Language Models}}, url = {https://github.com/ruanchaves/eplm}, version = {1.0.0}, year = {2023} } ```	null	Non_BioNLP	# BERTimbau base for Recognizing Textual Entailment This is the [neuralmind/bert-base-portuguese-cased](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model finetuned for Recognizing Textual Entailment with the [ASSIN](https://huggingface.co/datasets/assin) dataset. This model is suitable for Portuguese. - Git Repo: [Evaluation of Portuguese Language Models](https://github.com/ruanchaves/eplm). - Demo: [Portuguese Textual Entailment](https://ruanchaves-portuguese-textual-entailment.hf.space) ### Labels: * 0 : There is no entailment between premise and hypothesis. * 1 : There is entailment between premise and hypothesis. * 2 : The premise is a paraphrase of the hypothesis. ## Full classification example ```python from transformers import AutoModelForSequenceClassification, AutoTokenizer, AutoConfig import numpy as np import torch from scipy.special import softmax model_name = "ruanchaves/bert-base-portuguese-cased-assin-entailment" s1 = "Os homens estão cuidadosamente colocando as malas no porta-malas de um carro." s2 = "Os homens estão colocando bagagens dentro do porta-malas de um carro." model = AutoModelForSequenceClassification.from_pretrained(model_name) tokenizer = AutoTokenizer.from_pretrained(model_name) config = AutoConfig.from_pretrained(model_name) model_input = tokenizer(([s1], [s2]), padding=True, return_tensors="pt") with torch.no_grad(): output = model(*model_input) scores = output[0][0].detach().numpy() scores = softmax(scores) ranking = np.argsort(scores) ranking = ranking[::-1] for i in range(scores.shape[0]): l = config.id2label[ranking[i]] s = scores[ranking[i]] print(f"{i+1}) Label: {l} Score: {np.round(float(s), 4)}") ``` ## Citation Our research is ongoing, and we are currently working on describing our experiments in a paper, which will be published soon. In the meanwhile, if you would like to cite our work or models before the publication of the paper, please cite our [GitHub repository](https://github.com/ruanchaves/eplm): ``` @software{Chaves_Rodrigues_eplm_2023, author = {Chaves Rodrigues, Ruan and Tanti, Marc and Agerri, Rodrigo}, doi = {10.5281/zenodo.7781848}, month = {3}, title = {{Evaluation of Portuguese Language Models}}, url = {https://github.com/ruanchaves/eplm}, version = {1.0.0}, year = {2023} } ```	{"datasets": ["assin"], "language": "pt", "inference": false}	task	[ "TEXTUAL_ENTAILMENT" ]	44,555
neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic	neuralmagic	text-generation	[ "transformers", "safetensors", "qwen2", "text-generation", "deepseek", "fp8", "vllm", "conversational", "base_model:deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "base_model:quantized:deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "license:mit", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "compressed-tensors", "region:us" ]	2025-02-01T18:03:58Z	2025-02-27T08:10:20+00:00	1,069	1	--- base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B library_name: transformers license: mit tags: - deepseek - fp8 - vllm --- # DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic ## Model Overview - Model Architecture: Qwen2ForCausalLM - Input: Text - Output: Text - Model Optimizations: - Weight quantization: FP8 - Activation quantization: FP8 - Release Date: 2/5/2025 - Version: 1.0 - Model Developers: Neural Magic Quantized version of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B). ### Model Optimizations This model was obtained by quantizing the weights and activations of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B) to FP8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks are quantized. Weights are quantized using a symmetric per-channel scheme, whereas quantizations are quantized using a symmetric per-token scheme. [LLM Compressor](https://github.com/vllm-project/llm-compressor) is used for quantization. ## Use with vLLM This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams number_gpus = 1 model_name = "neuralmagic/DeepSeek-R1-Distill-Qwen-7B-dynamic" tokenizer = AutoTokenizer.from_pretrained(model_name) sampling_params = SamplingParams(temperature=0.6, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) llm = LLM(model=model_name, tensor_parallel_size=number_gpus, trust_remote_code=True) messages_list = [ [{"role": "user", "content": "Who are you? Please respond in pirate speak!"}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details. ## Creation This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below. ```python from transformers import AutoModelForCausalLM, AutoTokenizer from llmcompressor.modifiers.quantization import QuantizationModifier from llmcompressor.transformers import oneshot import os # Load model model_stub = "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B" model_name = model_stub.split("/")[-1] model = AutoModelForCausalLM.from_pretrained( model_stub, torch_dtype="auto", ) tokenizer = AutoTokenizer.from_pretrained(model_stub) # Configure the quantization algorithm and scheme recipe = QuantizationModifier( targets="Linear", scheme="FP8_DYNAMIC", ignore=["lm_head"], ) # Apply quantization oneshot( model=model, recipe=recipe, ) # Save to disk in compressed-tensors format save_path = model_name + "-FP8-dynamic model.save_pretrained(save_path) tokenizer.save_pretrained(save_path) print(f"Model and tokenizer saved to: {save_path}") ``` ## Evaluation The model was evaluated on OpenLLM Leaderboard [V1](https://huggingface.co/spaces/open-llm-leaderboard-old/open_llm_leaderboard) and [V2](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/), using the following commands: OpenLLM Leaderboard V1: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --tasks openllm \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` OpenLLM Leaderboard V2: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --apply_chat_template \ --fewshot_as_multiturn \ --tasks leaderboard \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` ### Accuracy <table> <thead> <tr> <th>Category</th> <th>Metric</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <th>Recovery</th> </tr> </thead> <tbody> <tr> <td rowspan="4"><b>Reasoning</b></td> <td>AIME 2024 (pass@1)</td> <td>53.17</td> <td>53.17</td> <td>100%</td> </tr> <tr> <td>MATH-500 (pass@1)</td> <td>93.66</td> <td>93.62</td> <td>99.96%</td> </tr> <tr> <td>GPQA Diamond (pass@1)</td> <td>50.53</td> <td>50.28</td> <td>99.51%</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>65.79</b></td> <td><b>65.69</b></td> <td><b>99.85%</b></td> </tr> <tr> <td rowspan="7"><b>OpenLLM V1</b></td> <td>ARC-Challenge (Acc-Norm, 25-shot)</td> <td>50.51</td> <td>50.51</td> <td>100.0%</td> </tr> <tr> <td>GSM8K (Strict-Match, 5-shot)</td> <td>78.62</td> <td>79.83</td> <td>101.5%</td> </tr> <tr> <td>HellaSwag (Acc-Norm, 10-shot)</td> <td>61.90</td> <td>61.62</td> <td>99.6%</td> </tr> <tr> <td>MMLU (Acc, 5-shot)</td> <td>54.19</td> <td>53.76</td> <td>99.2%</td> </tr> <tr> <td>TruthfulQA (MC2, 0-shot)</td> <td>45.55</td> <td>46.14</td> <td>101.3%</td> </tr> <tr> <td>Winogrande (Acc, 5-shot)</td> <td>61.56</td> <td>60.54</td> <td>98.3%</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>58.72</b></td> <td><b>58.73</b></td> <td><b>100.0%</b></td> </tr> <tr> <td rowspan="7"><b>OpenLLM V2</b></td> <td>IFEval (Inst Level Strict Acc, 0-shot)</td> <td>39.38</td> <td>39.01</td> <td>99.1%</td> </tr> <tr> <td>BBH (Acc-Norm, 3-shot)</td> <td>6.97</td> <td>6.19</td> <td>---</td> </tr> <tr> <td>Math-Hard (Exact-Match, 4-shot)</td> <td>0.00</td> <td>0.00</td> <td>---</td> </tr> <tr> <td>GPQA (Acc-Norm, 0-shot)</td> <td>1.81</td> <td>1.63</td> <td>---</td> </tr> <tr> <td>MUSR (Acc-Norm, 0-shot)</td> <td>4.68</td> <td>5.08</td> <td>---</td> </tr> <tr> <td>MMLU-Pro (Acc, 5-shot)</td> <td>1.66</td> <td>1.76</td> <td>---</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>9.08</b></td> <td><b>8.94</b></td> <td><b>---</b></td> </tr> <tr> <td rowspan="4"><b>Coding</b></td> <td>HumanEval (pass@1)</td> <td>40.80</td> <td>39.50</td> <td><b>96.8%</b></td> </tr> <tr> <td>HumanEval (pass@10)</td> <td>64.40</td> <td>62.10</td> <td>96.4%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>38.50</td> <td>37.20</td> <td>96.6%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>60.40</td> <td>59.30</td> <td>98.2%</td> </tr> </tbody> </table> ## Inference Performance This model achieves up to 1.4x speedup in single-stream deployment and up to 1.2x speedup in multi-stream asynchronous deployment, depending on hardware and use-case scenario. The following performance benchmarks were conducted with [vLLM](https://docs.vllm.ai/en/latest/) version 0.7.2, and [GuideLLM](https://github.com/neuralmagic/guidellm). <details> <summary>Benchmarking Command</summary> ``` guidellm --model neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic --target "http://localhost:8000/v1" --data-type emulated --data "prompt_tokens=<prompt_tokens>,generated_tokens=<generated_tokens>" --max seconds 360 --backend aiohttp_server ``` </details> ### Single-stream performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following<br>256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat<br>512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation<br>768 / 128</th> <th style="text-align: center;" colspan="2" >RAG<br>1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion<br>256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing<br>1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization<br>4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG<br>10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>2.9</td> <td>1576</td> <td>5.7</td> <td>788</td> <td>2.9</td> <td>1535</td> <td>3.0</td> <td>1496</td> <td>22.6</td> <td>199</td> <td>23.2</td> <td>194</td> <td>12.1</td> <td>370</td> <td>38.5</td> <td>117</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.56</td> <td>1.8</td> <td>2495</td> <td>3.7</td> <td>1223</td> <td>1.9</td> <td>2384</td> <td>1.9</td> <td>2393</td> <td>14.3</td> <td>315</td> <td>14.8</td> <td>304</td> <td>7.9</td> <td>572</td> <td>25.3</td> <td>178</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>2.41</td> <td>1.1</td> <td>4086</td> <td>2.3</td> <td>1998</td> <td>1.2</td> <td>3783</td> <td>1.3</td> <td>3527</td> <td>8.6</td> <td>526</td> <td>8.8</td> <td>512</td> <td>5.2</td> <td>860</td> <td>22.7</td> <td>198</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>1.4</td> <td>1389</td> <td>2.9</td> <td>691</td> <td>1.5</td> <td>1358</td> <td>1.5</td> <td>1329</td> <td>11.5</td> <td>175</td> <td>11.6</td> <td>174</td> <td>6.2</td> <td>326</td> <td>21.5</td> <td>93</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.28</td> <td>1.1</td> <td>1850</td> <td>2.2</td> <td>905</td> <td>1.1</td> <td>1807</td> <td>1.1</td> <td>1750</td> <td>8.6</td> <td>233</td> <td>8.7</td> <td>230</td> <td>4.7</td> <td>431</td> <td>23.1</td> <td>87</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.72</td> <td>0.8</td> <td>2575</td> <td>1.5</td> <td>1298</td> <td>0.8</td> <td>2461</td> <td>0.8</td> <td>2382</td> <td>6.1</td> <td>331</td> <td>6.2</td> <td>323</td> <td>3.6</td> <td>566</td> <td>22.7</td> <td>89</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>0.9</td> <td>1161</td> <td>1.9</td> <td>579</td> <td>1.0</td> <td>1138</td> <td>1.0</td> <td>1121</td> <td>7.5</td> <td>146</td> <td>7.6</td> <td>145</td> <td>3.9</td> <td>279</td> <td>15.4</td> <td>71</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.34</td> <td>0.7</td> <td>1585</td> <td>1.4</td> <td>786</td> <td>0.7</td> <td>1577</td> <td>0.7</td> <td>1524</td> <td>5.3</td> <td>207</td> <td>5.5</td> <td>197</td> <td>2.9</td> <td>382</td> <td>14.3</td> <td>77</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.33</td> <td>0.7</td> <td>1590</td> <td>1.4</td> <td>793</td> <td>0.7</td> <td>1549</td> <td>0.7</td> <td>1509</td> <td>5.4</td> <td>201</td> <td>5.5</td> <td>198</td> <td>2.9</td> <td>381</td> <td>14.0</td> <td>78</td> </tr> </tbody> </table> Use case profiles: prompt tokens / generation tokens QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025). ### Multi-stream asynchronous performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following<br>256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat<br>512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation<br>768 / 128</th> <th style="text-align: center;" colspan="2" >RAG<br>1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion<br>256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing<br>1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization<br>4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG<br>10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>14.9</td> <td>67138</td> <td>7.1</td> <td>32094</td> <td>7.4</td> <td>33096</td> <td>5.9</td> <td>26480</td> <td>2.0</td> <td>9004</td> <td>1.5</td> <td>6639</td> <td>1.1</td> <td>4938</td> <td>0.3</td> <td>1151</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.36</td> <td>20.2</td> <td>90956</td> <td>8.8</td> <td>39786</td> <td>10.2</td> <td>45963</td> <td>8.1</td> <td>36596</td> <td>3.1</td> <td>13968</td> <td>2.1</td> <td>9629</td> <td>1.4</td> <td>6374</td> <td>0.3</td> <td>1429</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.00</td> <td>13.3</td> <td>59681</td> <td>6.1</td> <td>27633</td> <td>5.9</td> <td>26689</td> <td>4.7</td> <td>20944</td> <td>2.9</td> <td>13108</td> <td>1.9</td> <td>8355</td> <td>1.0</td> <td>4362</td> <td>0.3</td> <td>1170</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>26.4</td> <td>53073</td> <td>13.0</td> <td>26213</td> <td>14.5</td> <td>29110</td> <td>11.4</td> <td>22936</td> <td>4.4</td> <td>8749</td> <td>3.3</td> <td>6680</td> <td>2.3</td> <td>4634</td> <td>0.5</td> <td>1105</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.27</td> <td>34.3</td> <td>69009</td> <td>14.8</td> <td>29791</td> <td>19.0</td> <td>38214</td> <td>15.7</td> <td>31598</td> <td>5.6</td> <td>11186</td> <td>4.2</td> <td>8350</td> <td>3.0</td> <td>6020</td> <td>0.7</td> <td>1328</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>0.93</td> <td>23.9</td> <td>47993</td> <td>12.0</td> <td>24194</td> <td>12.5</td> <td>25239</td> <td>10.0</td> <td>20029</td> <td>4.5</td> <td>9055</td> <td>3.3</td> <td>6681</td> <td>2.1</td> <td>4156</td> <td>0.5</td> <td>1043</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>54.3</td> <td>59410</td> <td>26.0</td> <td>28440</td> <td>32.1</td> <td>35154</td> <td>26.7</td> <td>29190</td> <td>8.0</td> <td>8700</td> <td>6.6</td> <td>7275</td> <td>5.2</td> <td>5669</td> <td>1.2</td> <td>1266</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.16</td> <td>62.9</td> <td>68818</td> <td>30.3</td> <td>33196</td> <td>39.4</td> <td>43132</td> <td>31.1</td> <td>34073</td> <td>9.2</td> <td>10058</td> <td>7.1</td> <td>7748</td> <td>6.1</td> <td>6714</td> <td>1.3</td> <td>1415</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.02</td> <td>56.2</td> <td>61483</td> <td>26.7</td> <td>29243</td> <td>32.5</td> <td>35592</td> <td>26.9</td> <td>29461</td> <td>8.3</td> <td>9072</td> <td>6.4</td> <td>7027</td> <td>5.2</td> <td>5731</td> <td>1.2</td> <td>1291</td> </tr> </tbody> </table> Use case profiles: prompt tokens / generation tokens QPS: Queries per second. **QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025).	null	Non_BioNLP	# DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic ## Model Overview - Model Architecture: Qwen2ForCausalLM - Input: Text - Output: Text - Model Optimizations: - Weight quantization: FP8 - Activation quantization: FP8 - Release Date: 2/5/2025 - Version: 1.0 - Model Developers: Neural Magic Quantized version of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B). ### Model Optimizations This model was obtained by quantizing the weights and activations of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B) to FP8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks are quantized. Weights are quantized using a symmetric per-channel scheme, whereas quantizations are quantized using a symmetric per-token scheme. [LLM Compressor](https://github.com/vllm-project/llm-compressor) is used for quantization. ## Use with vLLM This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams number_gpus = 1 model_name = "neuralmagic/DeepSeek-R1-Distill-Qwen-7B-dynamic" tokenizer = AutoTokenizer.from_pretrained(model_name) sampling_params = SamplingParams(temperature=0.6, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) llm = LLM(model=model_name, tensor_parallel_size=number_gpus, trust_remote_code=True) messages_list = [ [{"role": "user", "content": "Who are you? Please respond in pirate speak!"}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details. ## Creation This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below. ```python from transformers import AutoModelForCausalLM, AutoTokenizer from llmcompressor.modifiers.quantization import QuantizationModifier from llmcompressor.transformers import oneshot import os # Load model model_stub = "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B" model_name = model_stub.split("/")[-1] model = AutoModelForCausalLM.from_pretrained( model_stub, torch_dtype="auto", ) tokenizer = AutoTokenizer.from_pretrained(model_stub) # Configure the quantization algorithm and scheme recipe = QuantizationModifier( targets="Linear", scheme="FP8_DYNAMIC", ignore=["lm_head"], ) # Apply quantization oneshot( model=model, recipe=recipe, ) # Save to disk in compressed-tensors format save_path = model_name + "-FP8-dynamic model.save_pretrained(save_path) tokenizer.save_pretrained(save_path) print(f"Model and tokenizer saved to: {save_path}") ``` ## Evaluation The model was evaluated on OpenLLM Leaderboard [V1](https://huggingface.co/spaces/open-llm-leaderboard-old/open_llm_leaderboard) and [V2](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/), using the following commands: OpenLLM Leaderboard V1: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --tasks openllm \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` OpenLLM Leaderboard V2: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --apply_chat_template \ --fewshot_as_multiturn \ --tasks leaderboard \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` ### Accuracy <table> <thead> <tr> <th>Category</th> <th>Metric</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <th>Recovery</th> </tr> </thead> <tbody> <tr> <td rowspan="4"><b>Reasoning</b></td> <td>AIME 2024 (pass@1)</td> <td>53.17</td> <td>53.17</td> <td>100%</td> </tr> <tr> <td>MATH-500 (pass@1)</td> <td>93.66</td> <td>93.62</td> <td>99.96%</td> </tr> <tr> <td>GPQA Diamond (pass@1)</td> <td>50.53</td> <td>50.28</td> <td>99.51%</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>65.79</b></td> <td><b>65.69</b></td> <td><b>99.85%</b></td> </tr> <tr> <td rowspan="7"><b>OpenLLM V1</b></td> <td>ARC-Challenge (Acc-Norm, 25-shot)</td> <td>50.51</td> <td>50.51</td> <td>100.0%</td> </tr> <tr> <td>GSM8K (Strict-Match, 5-shot)</td> <td>78.62</td> <td>79.83</td> <td>101.5%</td> </tr> <tr> <td>HellaSwag (Acc-Norm, 10-shot)</td> <td>61.90</td> <td>61.62</td> <td>99.6%</td> </tr> <tr> <td>MMLU (Acc, 5-shot)</td> <td>54.19</td> <td>53.76</td> <td>99.2%</td> </tr> <tr> <td>TruthfulQA (MC2, 0-shot)</td> <td>45.55</td> <td>46.14</td> <td>101.3%</td> </tr> <tr> <td>Winogrande (Acc, 5-shot)</td> <td>61.56</td> <td>60.54</td> <td>98.3%</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>58.72</b></td> <td><b>58.73</b></td> <td><b>100.0%</b></td> </tr> <tr> <td rowspan="7"><b>OpenLLM V2</b></td> <td>IFEval (Inst Level Strict Acc, 0-shot)</td> <td>39.38</td> <td>39.01</td> <td>99.1%</td> </tr> <tr> <td>BBH (Acc-Norm, 3-shot)</td> <td>6.97</td> <td>6.19</td> <td>---</td> </tr> <tr> <td>Math-Hard (Exact-Match, 4-shot)</td> <td>0.00</td> <td>0.00</td> <td>---</td> </tr> <tr> <td>GPQA (Acc-Norm, 0-shot)</td> <td>1.81</td> <td>1.63</td> <td>---</td> </tr> <tr> <td>MUSR (Acc-Norm, 0-shot)</td> <td>4.68</td> <td>5.08</td> <td>---</td> </tr> <tr> <td>MMLU-Pro (Acc, 5-shot)</td> <td>1.66</td> <td>1.76</td> <td>---</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>9.08</b></td> <td><b>8.94</b></td> <td><b>---</b></td> </tr> <tr> <td rowspan="4"><b>Coding</b></td> <td>HumanEval (pass@1)</td> <td>40.80</td> <td>39.50</td> <td><b>96.8%</b></td> </tr> <tr> <td>HumanEval (pass@10)</td> <td>64.40</td> <td>62.10</td> <td>96.4%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>38.50</td> <td>37.20</td> <td>96.6%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>60.40</td> <td>59.30</td> <td>98.2%</td> </tr> </tbody> </table> ## Inference Performance This model achieves up to 1.4x speedup in single-stream deployment and up to 1.2x speedup in multi-stream asynchronous deployment, depending on hardware and use-case scenario. The following performance benchmarks were conducted with [vLLM](https://docs.vllm.ai/en/latest/) version 0.7.2, and [GuideLLM](https://github.com/neuralmagic/guidellm). <details> <summary>Benchmarking Command</summary> ``` guidellm --model neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic --target "http://localhost:8000/v1" --data-type emulated --data "prompt_tokens=<prompt_tokens>,generated_tokens=<generated_tokens>" --max seconds 360 --backend aiohttp_server ``` </details> ### Single-stream performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following<br>256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat<br>512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation<br>768 / 128</th> <th style="text-align: center;" colspan="2" >RAG<br>1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion<br>256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing<br>1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization<br>4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG<br>10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>2.9</td> <td>1576</td> <td>5.7</td> <td>788</td> <td>2.9</td> <td>1535</td> <td>3.0</td> <td>1496</td> <td>22.6</td> <td>199</td> <td>23.2</td> <td>194</td> <td>12.1</td> <td>370</td> <td>38.5</td> <td>117</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.56</td> <td>1.8</td> <td>2495</td> <td>3.7</td> <td>1223</td> <td>1.9</td> <td>2384</td> <td>1.9</td> <td>2393</td> <td>14.3</td> <td>315</td> <td>14.8</td> <td>304</td> <td>7.9</td> <td>572</td> <td>25.3</td> <td>178</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>2.41</td> <td>1.1</td> <td>4086</td> <td>2.3</td> <td>1998</td> <td>1.2</td> <td>3783</td> <td>1.3</td> <td>3527</td> <td>8.6</td> <td>526</td> <td>8.8</td> <td>512</td> <td>5.2</td> <td>860</td> <td>22.7</td> <td>198</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>1.4</td> <td>1389</td> <td>2.9</td> <td>691</td> <td>1.5</td> <td>1358</td> <td>1.5</td> <td>1329</td> <td>11.5</td> <td>175</td> <td>11.6</td> <td>174</td> <td>6.2</td> <td>326</td> <td>21.5</td> <td>93</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.28</td> <td>1.1</td> <td>1850</td> <td>2.2</td> <td>905</td> <td>1.1</td> <td>1807</td> <td>1.1</td> <td>1750</td> <td>8.6</td> <td>233</td> <td>8.7</td> <td>230</td> <td>4.7</td> <td>431</td> <td>23.1</td> <td>87</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.72</td> <td>0.8</td> <td>2575</td> <td>1.5</td> <td>1298</td> <td>0.8</td> <td>2461</td> <td>0.8</td> <td>2382</td> <td>6.1</td> <td>331</td> <td>6.2</td> <td>323</td> <td>3.6</td> <td>566</td> <td>22.7</td> <td>89</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>0.9</td> <td>1161</td> <td>1.9</td> <td>579</td> <td>1.0</td> <td>1138</td> <td>1.0</td> <td>1121</td> <td>7.5</td> <td>146</td> <td>7.6</td> <td>145</td> <td>3.9</td> <td>279</td> <td>15.4</td> <td>71</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.34</td> <td>0.7</td> <td>1585</td> <td>1.4</td> <td>786</td> <td>0.7</td> <td>1577</td> <td>0.7</td> <td>1524</td> <td>5.3</td> <td>207</td> <td>5.5</td> <td>197</td> <td>2.9</td> <td>382</td> <td>14.3</td> <td>77</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.33</td> <td>0.7</td> <td>1590</td> <td>1.4</td> <td>793</td> <td>0.7</td> <td>1549</td> <td>0.7</td> <td>1509</td> <td>5.4</td> <td>201</td> <td>5.5</td> <td>198</td> <td>2.9</td> <td>381</td> <td>14.0</td> <td>78</td> </tr> </tbody> </table> Use case profiles: prompt tokens / generation tokens QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025). ### Multi-stream asynchronous performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following<br>256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat<br>512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation<br>768 / 128</th> <th style="text-align: center;" colspan="2" >RAG<br>1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion<br>256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing<br>1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization<br>4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG<br>10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>14.9</td> <td>67138</td> <td>7.1</td> <td>32094</td> <td>7.4</td> <td>33096</td> <td>5.9</td> <td>26480</td> <td>2.0</td> <td>9004</td> <td>1.5</td> <td>6639</td> <td>1.1</td> <td>4938</td> <td>0.3</td> <td>1151</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.36</td> <td>20.2</td> <td>90956</td> <td>8.8</td> <td>39786</td> <td>10.2</td> <td>45963</td> <td>8.1</td> <td>36596</td> <td>3.1</td> <td>13968</td> <td>2.1</td> <td>9629</td> <td>1.4</td> <td>6374</td> <td>0.3</td> <td>1429</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.00</td> <td>13.3</td> <td>59681</td> <td>6.1</td> <td>27633</td> <td>5.9</td> <td>26689</td> <td>4.7</td> <td>20944</td> <td>2.9</td> <td>13108</td> <td>1.9</td> <td>8355</td> <td>1.0</td> <td>4362</td> <td>0.3</td> <td>1170</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>26.4</td> <td>53073</td> <td>13.0</td> <td>26213</td> <td>14.5</td> <td>29110</td> <td>11.4</td> <td>22936</td> <td>4.4</td> <td>8749</td> <td>3.3</td> <td>6680</td> <td>2.3</td> <td>4634</td> <td>0.5</td> <td>1105</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.27</td> <td>34.3</td> <td>69009</td> <td>14.8</td> <td>29791</td> <td>19.0</td> <td>38214</td> <td>15.7</td> <td>31598</td> <td>5.6</td> <td>11186</td> <td>4.2</td> <td>8350</td> <td>3.0</td> <td>6020</td> <td>0.7</td> <td>1328</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>0.93</td> <td>23.9</td> <td>47993</td> <td>12.0</td> <td>24194</td> <td>12.5</td> <td>25239</td> <td>10.0</td> <td>20029</td> <td>4.5</td> <td>9055</td> <td>3.3</td> <td>6681</td> <td>2.1</td> <td>4156</td> <td>0.5</td> <td>1043</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>54.3</td> <td>59410</td> <td>26.0</td> <td>28440</td> <td>32.1</td> <td>35154</td> <td>26.7</td> <td>29190</td> <td>8.0</td> <td>8700</td> <td>6.6</td> <td>7275</td> <td>5.2</td> <td>5669</td> <td>1.2</td> <td>1266</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.16</td> <td>62.9</td> <td>68818</td> <td>30.3</td> <td>33196</td> <td>39.4</td> <td>43132</td> <td>31.1</td> <td>34073</td> <td>9.2</td> <td>10058</td> <td>7.1</td> <td>7748</td> <td>6.1</td> <td>6714</td> <td>1.3</td> <td>1415</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.02</td> <td>56.2</td> <td>61483</td> <td>26.7</td> <td>29243</td> <td>32.5</td> <td>35592</td> <td>26.9</td> <td>29461</td> <td>8.3</td> <td>9072</td> <td>6.4</td> <td>7027</td> <td>5.2</td> <td>5731</td> <td>1.2</td> <td>1291</td> </tr> </tbody> </table> Use case profiles: prompt tokens / generation tokens QPS: Queries per second. **QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025).	{"base_model": "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "library_name": "transformers", "license": "mit", "tags": ["deepseek", "fp8", "vllm"]}	task	[ "SUMMARIZATION" ]	44,556
AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-product-reviews	AdapterHub	null	[ "adapter-transformers", "xlm-roberta", "dataset:UKPLab/m2qa", "region:us" ]	2024-06-05T20:09:06Z	2024-12-11T11:28:56+00:00	8	0	--- datasets: - UKPLab/m2qa tags: - adapter-transformers - xlm-roberta --- # M2QA Adapter: Domain Adapter for MAD-X+Domain Setup This adapter is part of the M2QA publication to achieve language and domain transfer via adapters. 📃 Paper: [https://aclanthology.org/2024.findings-emnlp.365/](https://aclanthology.org/2024.findings-emnlp.365/) 🏗️ GitHub repo: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa) 💾 Hugging Face Dataset: [https://huggingface.co/UKPLab/m2qa](https://huggingface.co/UKPLab/m2qa) Important: This adapter only works together with the MAD-X language adapters and the M2QA QA head adapter. This [adapter](https://adapterhub.ml) for the `xlm-roberta-base` model that was trained using the [Adapters](https://github.com/Adapter-Hub/adapters) library. For detailed training details see our paper or GitHub repository: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa). You can find the evaluation results for this adapter on the M2QA dataset in the GitHub repo and in the paper. ## Usage First, install `adapters`: ``` pip install -U adapters ``` Now, the adapter can be loaded and activated like this: ```python from adapters import AutoAdapterModel from adapters.composition import Stack model = AutoAdapterModel.from_pretrained("xlm-roberta-base") # 1. Load language adapter language_adapter_name = model.load_adapter("de/wiki@ukp") # MAD-X+Domain uses the MAD-X language adapter # 2. Load domain adapter domain_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-product-reviews") # 3. Load QA head adapter qa_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-qa-head") # 4. Activate them via the adapter stack model.active_adapters = Stack(language_adapter_name, domain_adapter_name, qa_adapter_name) ``` See our repository for more information: See https://github.com/UKPLab/m2qa/tree/main/Experiments/mad-x-domain ## Contact Leon Engländer: - [HuggingFace Profile](https://huggingface.co/lenglaender) - [GitHub](https://github.com/lenglaender) - [Twitter](https://x.com/LeonEnglaender) ## Citation ``` @inproceedings{englander-etal-2024-m2qa, title = "M2QA: Multi-domain Multilingual Question Answering", author = {Engl{\"a}nder, Leon and Sterz, Hannah and Poth, Clifton A and Pfeiffer, Jonas and Kuznetsov, Ilia and Gurevych, Iryna}, booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024", month = nov, year = "2024", address = "Miami, Florida, USA", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.findings-emnlp.365", pages = "6283--6305", } ```	null	Non_BioNLP	# M2QA Adapter: Domain Adapter for MAD-X+Domain Setup This adapter is part of the M2QA publication to achieve language and domain transfer via adapters. 📃 Paper: [https://aclanthology.org/2024.findings-emnlp.365/](https://aclanthology.org/2024.findings-emnlp.365/) 🏗️ GitHub repo: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa) 💾 Hugging Face Dataset: [https://huggingface.co/UKPLab/m2qa](https://huggingface.co/UKPLab/m2qa) Important: This adapter only works together with the MAD-X language adapters and the M2QA QA head adapter. This [adapter](https://adapterhub.ml) for the `xlm-roberta-base` model that was trained using the [Adapters](https://github.com/Adapter-Hub/adapters) library. For detailed training details see our paper or GitHub repository: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa). You can find the evaluation results for this adapter on the M2QA dataset in the GitHub repo and in the paper. ## Usage First, install `adapters`: ``` pip install -U adapters ``` Now, the adapter can be loaded and activated like this: ```python from adapters import AutoAdapterModel from adapters.composition import Stack model = AutoAdapterModel.from_pretrained("xlm-roberta-base") # 1. Load language adapter language_adapter_name = model.load_adapter("de/wiki@ukp") # MAD-X+Domain uses the MAD-X language adapter # 2. Load domain adapter domain_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-product-reviews") # 3. Load QA head adapter qa_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-qa-head") # 4. Activate them via the adapter stack model.active_adapters = Stack(language_adapter_name, domain_adapter_name, qa_adapter_name) ``` See our repository for more information: See https://github.com/UKPLab/m2qa/tree/main/Experiments/mad-x-domain ## Contact Leon Engländer: - [HuggingFace Profile](https://huggingface.co/lenglaender) - [GitHub](https://github.com/lenglaender) - [Twitter](https://x.com/LeonEnglaender) ## Citation ``` @inproceedings{englander-etal-2024-m2qa, title = "M2QA: Multi-domain Multilingual Question Answering", author = {Engl{\"a}nder, Leon and Sterz, Hannah and Poth, Clifton A and Pfeiffer, Jonas and Kuznetsov, Ilia and Gurevych, Iryna}, booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024", month = nov, year = "2024", address = "Miami, Florida, USA", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.findings-emnlp.365", pages = "6283--6305", } ```	{"datasets": ["UKPLab/m2qa"], "tags": ["adapter-transformers", "xlm-roberta"]}	task	[ "QUESTION_ANSWERING" ]	44,557
HelgeKn/SemEval-multi-class-4	HelgeKn	text-classification	[ "setfit", "safetensors", "mpnet", "sentence-transformers", "text-classification", "generated_from_setfit_trainer", "arxiv:2209.11055", "base_model:sentence-transformers/paraphrase-mpnet-base-v2", "base_model:finetune:sentence-transformers/paraphrase-mpnet-base-v2", "model-index", "region:us" ]	2023-12-14T13:05:58Z	2023-12-14T13:07:17+00:00	49	0	--- base_model: sentence-transformers/paraphrase-mpnet-base-v2 library_name: setfit metrics: - accuracy pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 'The Alavas worked themselves to the bone in the last period , and English and San Emeterio ( 65-75 ) had already made it clear that they were not going to let anyone take away what they had earned during the first thirty minutes . ' - text: 'To break the uncomfortable silence , Haney began to talk . ' - text: 'For the treatment of non-small cell lung cancer , the effects of Alimta were compared with those of docetaxel ( another anticancer medicine ) in one study involving 571 patients with locally advanced or metastatic disease who had received chemotherapy in the past . ' - text: 'As we all know , a few minutes before the end of the game ( that their team had already won ) , both players deliberately wasted time which made the referee show the second yellow card to both of them . ' - text: 'In contrast , patients whose cancer was affecting squamous cells had shorter survival times if they received Alimta . ' inference: true model-index: - name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2 results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: accuracy value: 0.14172185430463577 name: Accuracy --- # SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - Model Type: SetFit - Sentence Transformer body: [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - Classification head: a [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance - Maximum Sequence Length: 512 tokens - Number of Classes: 7 classes <!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Repository: [SetFit on GitHub](https://github.com/huggingface/setfit) - Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels \| Label \| Examples \| \|:------\|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| 1 \| <ul><li>'Eventually little French farmers and their little French farmwives came out of their stone houses and put their hands above their tiny eyes and squinted at us . '</li><li>'Mr. Neuberger realized that , although of Italian ancestry , Mr. Mariotta still could qualify as a minority person since he was born in Puerto Rico . '</li><li>"Biggest trouble was scared family who could n't get a phone line through , and spent a really horrible hour not knowing . "</li></ul> \| \| 4 \| <ul><li>'`` To ring for even one service at this tower , we have to scrape , `` says Mr. Hammond , a retired water-authority worker . `` '</li><li>"`` It 's my line of work `` , he said "</li><li>'One writer , signing his letter as `` Red-blooded , balanced male , `` remarked on the `` frequency of women fainting in peals , `` and suggested that they `` settle back into their traditional role of making tea at meetings . `` '</li></ul> \| \| 5 \| <ul><li>'Of course On Thursday , Haney mailed the monthly check for separate maintenance to his wife Lolly , and wished the stranger could do something about her '</li><li>"On the Right , the tone was set by Jacques Chirac , who declared in 1976 that `` 900,000 unemployed would not become a problem in a country with 2 million of foreign workers , '' and on the Left by Michel Rocard explaining in 1990 that France `` can not accommodate all the world 's misery . '' "</li><li>"But the council 's program to attract and train ringers is only partly successful , says Mr. Baldwin . "</li></ul> \| \| 6 \| <ul><li>'3 -RRB- Republican congressional representatives , because of their belief in a minimalist state , are less willing to engage in local benefit-seeking than are Democratic members of Congress . '</li><li>'As we know , voters tend to favor Republicans more in races for president than in those for Congress . '</li><li>'That is the way the system works . '</li></ul> \| \| 2 \| <ul><li>'-- Students should move up the educational ladder as their academic potential allows . '</li><li>'The next day , Sunday , the hangover reminded Haney where he had been the night before . '</li><li>'-- In most states , increasing expenditures on education , in our current circumstances , will probably make things worse , not better . '</li></ul> \| \| 0 \| <ul><li>'Then your focus will go to an input text box where you can type your function . '</li><li>"I might have got hit by that truck if it was n't for you . "</li><li>"Second , it explains why voters hold Congress in disdain but generally love their own congressional representatives : Any individual legislator 's constituents appreciate the specific benefits that the legislator wins for them but not the overall cost associated with every other legislator doing likewise for his own constituency . "</li></ul> \| \| 3 \| <ul><li>"It was the most exercise we 'd had all morning and it was followed by our driving immediately to the nearest watering hole . "</li><li>'Alimta is used together with cisplatin ( another anticancer medicine ) when the cancer is unresectable ( cannot be removed by surgery alone ) and malignant ( has spread , or is likely to spread easily , to other parts of the body ) , in patients who have not received chemotherapy ( medicines for cancer ) before advanced or metastatic non-small cell lung cancer that is not affecting the squamous cells . '</li><li>'If it is , it will be treated as an operator , if it is not , it will be treated as a user function . '</li></ul> \| ## Evaluation ### Metrics \| Label \| Accuracy \| \|:--------\|:---------\| \| all \| 0.1417 \| ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("HelgeKn/SemEval-multi-class-4") # Run inference preds = model("To break the uncomfortable silence , Haney began to talk . ") ``` <!-- ### Downstream Use List how someone could finetune this model on their own dataset. --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Set Metrics \| Training set \| Min \| Median \| Max \| \|:-------------\|:----\|:--------\|:----\| \| Word count \| 4 \| 27.1786 \| 74 \| \| Label \| Training Sample Count \| \|:------\|:----------------------\| \| 0 \| 4 \| \| 1 \| 4 \| \| 2 \| 4 \| \| 3 \| 4 \| \| 4 \| 4 \| \| 5 \| 4 \| \| 6 \| 4 \| ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (2, 2) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results \| Epoch \| Step \| Training Loss \| Validation Loss \| \|:------:\|:----:\|:-------------:\|:---------------:\| \| 0.0143 \| 1 \| 0.2446 \| - \| \| 0.7143 \| 50 \| 0.0612 \| - \| \| 1.4286 \| 100 \| 0.0078 \| - \| ### Framework Versions - Python: 3.9.13 - SetFit: 1.0.1 - Sentence Transformers: 2.2.2 - Transformers: 4.36.0 - PyTorch: 2.1.1+cpu - Datasets: 2.15.0 - Tokenizers: 0.15.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - Model Type: SetFit - Sentence Transformer body: [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - Classification head: a [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance - Maximum Sequence Length: 512 tokens - Number of Classes: 7 classes <!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Repository: [SetFit on GitHub](https://github.com/huggingface/setfit) - Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels \| Label \| Examples \| \|:------\|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| 1 \| <ul><li>'Eventually little French farmers and their little French farmwives came out of their stone houses and put their hands above their tiny eyes and squinted at us . '</li><li>'Mr. Neuberger realized that , although of Italian ancestry , Mr. Mariotta still could qualify as a minority person since he was born in Puerto Rico . '</li><li>"Biggest trouble was scared family who could n't get a phone line through , and spent a really horrible hour not knowing . "</li></ul> \| \| 4 \| <ul><li>'`` To ring for even one service at this tower , we have to scrape , `` says Mr. Hammond , a retired water-authority worker . `` '</li><li>"`` It 's my line of work `` , he said "</li><li>'One writer , signing his letter as `` Red-blooded , balanced male , `` remarked on the `` frequency of women fainting in peals , `` and suggested that they `` settle back into their traditional role of making tea at meetings . `` '</li></ul> \| \| 5 \| <ul><li>'Of course On Thursday , Haney mailed the monthly check for separate maintenance to his wife Lolly , and wished the stranger could do something about her '</li><li>"On the Right , the tone was set by Jacques Chirac , who declared in 1976 that `` 900,000 unemployed would not become a problem in a country with 2 million of foreign workers , '' and on the Left by Michel Rocard explaining in 1990 that France `` can not accommodate all the world 's misery . '' "</li><li>"But the council 's program to attract and train ringers is only partly successful , says Mr. Baldwin . "</li></ul> \| \| 6 \| <ul><li>'3 -RRB- Republican congressional representatives , because of their belief in a minimalist state , are less willing to engage in local benefit-seeking than are Democratic members of Congress . '</li><li>'As we know , voters tend to favor Republicans more in races for president than in those for Congress . '</li><li>'That is the way the system works . '</li></ul> \| \| 2 \| <ul><li>'-- Students should move up the educational ladder as their academic potential allows . '</li><li>'The next day , Sunday , the hangover reminded Haney where he had been the night before . '</li><li>'-- In most states , increasing expenditures on education , in our current circumstances , will probably make things worse , not better . '</li></ul> \| \| 0 \| <ul><li>'Then your focus will go to an input text box where you can type your function . '</li><li>"I might have got hit by that truck if it was n't for you . "</li><li>"Second , it explains why voters hold Congress in disdain but generally love their own congressional representatives : Any individual legislator 's constituents appreciate the specific benefits that the legislator wins for them but not the overall cost associated with every other legislator doing likewise for his own constituency . "</li></ul> \| \| 3 \| <ul><li>"It was the most exercise we 'd had all morning and it was followed by our driving immediately to the nearest watering hole . "</li><li>'Alimta is used together with cisplatin ( another anticancer medicine ) when the cancer is unresectable ( cannot be removed by surgery alone ) and malignant ( has spread , or is likely to spread easily , to other parts of the body ) , in patients who have not received chemotherapy ( medicines for cancer ) before advanced or metastatic non-small cell lung cancer that is not affecting the squamous cells . '</li><li>'If it is , it will be treated as an operator , if it is not , it will be treated as a user function . '</li></ul> \| ## Evaluation ### Metrics \| Label \| Accuracy \| \|:--------\|:---------\| \| all \| 0.1417 \| ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("HelgeKn/SemEval-multi-class-4") # Run inference preds = model("To break the uncomfortable silence , Haney began to talk . ") ``` <!-- ### Downstream Use List how someone could finetune this model on their own dataset. --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Set Metrics \| Training set \| Min \| Median \| Max \| \|:-------------\|:----\|:--------\|:----\| \| Word count \| 4 \| 27.1786 \| 74 \| \| Label \| Training Sample Count \| \|:------\|:----------------------\| \| 0 \| 4 \| \| 1 \| 4 \| \| 2 \| 4 \| \| 3 \| 4 \| \| 4 \| 4 \| \| 5 \| 4 \| \| 6 \| 4 \| ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (2, 2) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results \| Epoch \| Step \| Training Loss \| Validation Loss \| \|:------:\|:----:\|:-------------:\|:---------------:\| \| 0.0143 \| 1 \| 0.2446 \| - \| \| 0.7143 \| 50 \| 0.0612 \| - \| \| 1.4286 \| 100 \| 0.0078 \| - \| ### Framework Versions - Python: 3.9.13 - SetFit: 1.0.1 - Sentence Transformers: 2.2.2 - Transformers: 4.36.0 - PyTorch: 2.1.1+cpu - Datasets: 2.15.0 - Tokenizers: 0.15.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "sentence-transformers/paraphrase-mpnet-base-v2", "library_name": "setfit", "metrics": ["accuracy"], "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification", "generated_from_setfit_trainer"], "widget": [{"text": "The Alavas worked themselves to the bone in the last period , and English and San Emeterio ( 65-75 ) had already made it clear that they were not going to let anyone take away what they had earned during the first thirty minutes . "}, {"text": "To break the uncomfortable silence , Haney began to talk . "}, {"text": "For the treatment of non-small cell lung cancer , the effects of Alimta were compared with those of docetaxel ( another anticancer medicine ) in one study involving 571 patients with locally advanced or metastatic disease who had received chemotherapy in the past . "}, {"text": "As we all know , a few minutes before the end of the game ( that their team had already won ) , both players deliberately wasted time which made the referee show the second yellow card to both of them . "}, {"text": "In contrast , patients whose cancer was affecting squamous cells had shorter survival times if they received Alimta . "}], "inference": true, "model-index": [{"name": "SetFit with sentence-transformers/paraphrase-mpnet-base-v2", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "Unknown", "type": "unknown", "split": "test"}, "metrics": [{"type": "accuracy", "value": 0.14172185430463577, "name": "Accuracy"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,559
azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0	azale-ai	text-generation	[ "transformers", "safetensors", "mixtral", "text-generation", "merge", "mergekit", "lazymergekit", "moe", "indonesian", "multilingual", "en", "id", "jv", "su", "ms", "arxiv:2310.06825", "license:cc-by-nc-nd-4.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2024-01-13T06:00:11Z	2024-01-14T04:49:20+00:00	10	2	--- language: - en - id - jv - su - ms license: cc-by-nc-nd-4.0 tags: - merge - mergekit - lazymergekit - moe - indonesian - multilingual --- ![Gotong Royong Mixtral](./thumbnail.jpeg "Gotong Royong Mixtral: Generated by Bing AI Image Creator or Dall-E 3") # GotongRoyong-MixtralMoE-7Bx4-v1.0 GotongRoyong is a series of language models focused on Mixture of Experts (MoE), made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing) and [cg123/mergekit](https://github.com/cg123/mergekit). GotongRoyong-MixtralMoE-7Bx4-v1.0 is a specific variant of the open-source GotongRoyong language model that combines the architectural model [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1), but uses the base model from the specific fine-tuned version [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) with experts from [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha), [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4), [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2), and [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1). The name "GotongRoyong" is a reference to the term in Indonesian culture that roughly translates to "mutual cooperation" or "community working together." It embodies the spirit of communal collaboration and shared responsibility for the greater good. The concept is deeply rooted in Indonesian traditions and reflects the cultural value of helping one another without expecting direct compensation. ## Model Details - Model Size: 7Bx4 - Architecture Model: Mixure of Experts (MoE) using [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) - MoE Base Model: [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) - Expert Models: - [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha) - [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4) - [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2) - [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1) - License: [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id) ## How to use #### Installation To use GotongRoyong model, ensure that PyTorch has been installed and that you have an Nvidia GPU (or use Google Colab). After that you need to install the required dependencies: ```bash pip3 install -U bitsandbytes transformers peft accelerate einops evaluate scikit-learn sentencepiece ``` #### Usage Quantized Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", load_in_4bit=True, torch_dtype=torch.float32, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` #### Usage Normal Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", torch_dtype=torch.float16, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` ## Limitations 1. Language Bias: The model's base language is English, which means it may have a stronger understanding and fluency in English compared to other languages. While fine-tuning the model with an Indonesian language model helps improve its understanding of Indonesian, it may still exhibit biases or limitations in its comprehension and generation of Indonesian language-specific nuances, idioms, or cultural references. 2. Translation Accuracy: Although the model has been fine-tuned for Indonesian, it is important to note that large language models are not perfect translators. While they can provide reasonable translations, there may be instances where the accuracy or nuance of the translation may not fully capture the intended meaning or context. 3. Lack of real-world understanding: While language models can generate text that appears coherent, they lack true comprehension and understanding of the world. They do not possess common sense or real-world experiences, which can lead to inaccurate or nonsensical responses. 4. Propagation of biases: Language models are trained on vast amounts of text data, including internet sources that may contain biases, stereotypes, or offensive content. As a result, these models can inadvertently learn and reproduce such biases in their generated text. Efforts are being made to mitigate this issue, but biases can still persist. 5. Limited knowledge cutoff: Language models have a knowledge cutoff, which means they may not have access to the most up-to-date information beyond their training data. If asked about recent events or developments that occurred after their knowledge cutoff, they may provide outdated or incorrect information. 6. Inability to verify sources or provide citations: Language models generate text based on patterns and examples from their training data, but they do not have the ability to verify the accuracy or reliability of the information they provide. They cannot cite sources or provide evidence to support their claims. 7. Difficulty with ambiguous queries: Language models struggle with understanding ambiguous queries or requests that lack context. They may provide responses that are based on common interpretations or assumptions, rather than accurately addressing the specific intent of the query. 8. Ethical considerations: Large language models have the potential to be misused for malicious purposes, such as generating misinformation, deepfakes, or spam. Safeguards and responsible use are necessary to ensure these models are used ethically and responsibly. 9. Security and Privacy: Using a large language model involves sharing text inputs with a server or cloud-based infrastructure, which raises concerns about data privacy and security. Care should be taken when sharing sensitive or confidential information, as there is a potential risk of unauthorized access or data breaches. ## License The model is licensed under the [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id). ## Contributing We welcome contributions to enhance and improve our model. If you have any suggestions or find any issues, please feel free to open an issue or submit a pull request. Also we're open to sponsor for compute power. ## Contact Us For any further questions or assistance, please feel free to contact us using the information provided below. [[email protected]](mailto:[email protected]) ## Cite This Project ``` @software{Hafidh_Soekma_GotongRoyong_MixtralMoE_7Bx4_v1.0_2023, author = {Hafidh Soekma Ardiansyah}, month = january, title = {GotongRoyong: Indonesian Mixture Of Experts Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-beta}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2024} } ``` ## Citation ``` @software{Hafidh_Soekma_Startstreak_7b_alpha_2023, author = {Hafidh Soekma Ardiansyah}, month = october, title = {Startstreak: Traditional Indonesian Multilingual Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-alpha}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2023} } ``` ``` @article{Merak, title={Merak-7B: The LLM for Bahasa Indonesia}, author={Muhammad Ichsan}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @article{Sidrap, title={Sidrap-7B-v2: LLM Model for Bahasa Indonesia Dialog}, author={Robin Syihab}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @misc{Obrolin/Kesehatan-7B, author = {Arkan Bima}, title = {Obrolin Kesehatan}, publisher = {Hugging Face}, howpublished = {\url{https://huggingface.co/Obrolin/Kesehatan-7B}}, version = {0.1}, year = {2024}, } ``` ``` @misc{2310.06825, Author = {Albert Q. Jiang and Alexandre Sablayrolles and Arthur Mensch and Chris Bamford and Devendra Singh Chaplot and Diego de las Casas and Florian Bressand and Gianna Lengyel and Guillaume Lample and Lucile Saulnier and Lélio Renard Lavaud and Marie-Anne Lachaux and Pierre Stock and Teven Le Scao and Thibaut Lavril and Thomas Wang and Timothée Lacroix and William El Sayed}, Title = {Mistral 7B}, Year = {2023}, Eprint = {arXiv:2310.06825}, } ```	null	Non_BioNLP	![Gotong Royong Mixtral](./thumbnail.jpeg "Gotong Royong Mixtral: Generated by Bing AI Image Creator or Dall-E 3") # GotongRoyong-MixtralMoE-7Bx4-v1.0 GotongRoyong is a series of language models focused on Mixture of Experts (MoE), made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing) and [cg123/mergekit](https://github.com/cg123/mergekit). GotongRoyong-MixtralMoE-7Bx4-v1.0 is a specific variant of the open-source GotongRoyong language model that combines the architectural model [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1), but uses the base model from the specific fine-tuned version [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) with experts from [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha), [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4), [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2), and [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1). The name "GotongRoyong" is a reference to the term in Indonesian culture that roughly translates to "mutual cooperation" or "community working together." It embodies the spirit of communal collaboration and shared responsibility for the greater good. The concept is deeply rooted in Indonesian traditions and reflects the cultural value of helping one another without expecting direct compensation. ## Model Details - Model Size: 7Bx4 - Architecture Model: Mixure of Experts (MoE) using [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) - MoE Base Model: [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) - Expert Models: - [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha) - [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4) - [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2) - [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1) - License: [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id) ## How to use #### Installation To use GotongRoyong model, ensure that PyTorch has been installed and that you have an Nvidia GPU (or use Google Colab). After that you need to install the required dependencies: ```bash pip3 install -U bitsandbytes transformers peft accelerate einops evaluate scikit-learn sentencepiece ``` #### Usage Quantized Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", load_in_4bit=True, torch_dtype=torch.float32, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` #### Usage Normal Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", torch_dtype=torch.float16, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` ## Limitations 1. Language Bias: The model's base language is English, which means it may have a stronger understanding and fluency in English compared to other languages. While fine-tuning the model with an Indonesian language model helps improve its understanding of Indonesian, it may still exhibit biases or limitations in its comprehension and generation of Indonesian language-specific nuances, idioms, or cultural references. 2. Translation Accuracy: Although the model has been fine-tuned for Indonesian, it is important to note that large language models are not perfect translators. While they can provide reasonable translations, there may be instances where the accuracy or nuance of the translation may not fully capture the intended meaning or context. 3. Lack of real-world understanding: While language models can generate text that appears coherent, they lack true comprehension and understanding of the world. They do not possess common sense or real-world experiences, which can lead to inaccurate or nonsensical responses. 4. Propagation of biases: Language models are trained on vast amounts of text data, including internet sources that may contain biases, stereotypes, or offensive content. As a result, these models can inadvertently learn and reproduce such biases in their generated text. Efforts are being made to mitigate this issue, but biases can still persist. 5. Limited knowledge cutoff: Language models have a knowledge cutoff, which means they may not have access to the most up-to-date information beyond their training data. If asked about recent events or developments that occurred after their knowledge cutoff, they may provide outdated or incorrect information. 6. Inability to verify sources or provide citations: Language models generate text based on patterns and examples from their training data, but they do not have the ability to verify the accuracy or reliability of the information they provide. They cannot cite sources or provide evidence to support their claims. 7. Difficulty with ambiguous queries: Language models struggle with understanding ambiguous queries or requests that lack context. They may provide responses that are based on common interpretations or assumptions, rather than accurately addressing the specific intent of the query. 8. Ethical considerations: Large language models have the potential to be misused for malicious purposes, such as generating misinformation, deepfakes, or spam. Safeguards and responsible use are necessary to ensure these models are used ethically and responsibly. 9. Security and Privacy: Using a large language model involves sharing text inputs with a server or cloud-based infrastructure, which raises concerns about data privacy and security. Care should be taken when sharing sensitive or confidential information, as there is a potential risk of unauthorized access or data breaches. ## License The model is licensed under the [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id). ## Contributing We welcome contributions to enhance and improve our model. If you have any suggestions or find any issues, please feel free to open an issue or submit a pull request. Also we're open to sponsor for compute power. ## Contact Us For any further questions or assistance, please feel free to contact us using the information provided below. [[email protected]](mailto:[email protected]) ## Cite This Project ``` @software{Hafidh_Soekma_GotongRoyong_MixtralMoE_7Bx4_v1.0_2023, author = {Hafidh Soekma Ardiansyah}, month = january, title = {GotongRoyong: Indonesian Mixture Of Experts Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-beta}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2024} } ``` ## Citation ``` @software{Hafidh_Soekma_Startstreak_7b_alpha_2023, author = {Hafidh Soekma Ardiansyah}, month = october, title = {Startstreak: Traditional Indonesian Multilingual Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-alpha}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2023} } ``` ``` @article{Merak, title={Merak-7B: The LLM for Bahasa Indonesia}, author={Muhammad Ichsan}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @article{Sidrap, title={Sidrap-7B-v2: LLM Model for Bahasa Indonesia Dialog}, author={Robin Syihab}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @misc{Obrolin/Kesehatan-7B, author = {Arkan Bima}, title = {Obrolin Kesehatan}, publisher = {Hugging Face}, howpublished = {\url{https://huggingface.co/Obrolin/Kesehatan-7B}}, version = {0.1}, year = {2024}, } ``` ``` @misc{2310.06825, Author = {Albert Q. Jiang and Alexandre Sablayrolles and Arthur Mensch and Chris Bamford and Devendra Singh Chaplot and Diego de las Casas and Florian Bressand and Gianna Lengyel and Guillaume Lample and Lucile Saulnier and Lélio Renard Lavaud and Marie-Anne Lachaux and Pierre Stock and Teven Le Scao and Thibaut Lavril and Thomas Wang and Timothée Lacroix and William El Sayed}, Title = {Mistral 7B}, Year = {2023}, Eprint = {arXiv:2310.06825}, } ```	{"language": ["en", "id", "jv", "su", "ms"], "license": "cc-by-nc-nd-4.0", "tags": ["merge", "mergekit", "lazymergekit", "moe", "indonesian", "multilingual"]}	task	[ "TRANSLATION" ]	44,560
gokhalevikrant/finetuning-sentiment-model-Test	gokhalevikrant	text-classification	[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:imdb", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-12-06T13:42:39Z	2022-12-06T13:57:16+00:00	120	0	--- datasets: - imdb license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: finetuning-sentiment-model-Test results: - task: type: text-classification name: Text Classification dataset: name: imdb type: imdb config: plain_text split: train args: plain_text metrics: - type: accuracy value: 0.902 name: Accuracy - type: f1 value: 0.9037328094302554 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # finetuning-sentiment-model-Test This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2793 - Accuracy: 0.902 - F1: 0.9037 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.25.1 - Pytorch 1.12.1+cu113 - Datasets 2.7.1 - Tokenizers 0.13.2	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # finetuning-sentiment-model-Test This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2793 - Accuracy: 0.902 - F1: 0.9037 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.25.1 - Pytorch 1.12.1+cu113 - Datasets 2.7.1 - Tokenizers 0.13.2	{"datasets": ["imdb"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "finetuning-sentiment-model-Test", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "imdb", "type": "imdb", "config": "plain_text", "split": "train", "args": "plain_text"}, "metrics": [{"type": "accuracy", "value": 0.902, "name": "Accuracy"}, {"type": "f1", "value": 0.9037328094302554, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,561
michaellee8/wisper-base-jwt	michaellee8	automatic-speech-recognition	[ "transformers", "pytorch", "tf", "jax", "safetensors", "whisper", "automatic-speech-recognition", "audio", "hf-asr-leaderboard", "en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", "no", "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su", "arxiv:2212.04356", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]	2024-01-05T21:20:05Z	2024-01-06T12:17:00+00:00	14	0	--- language: - en - zh - de - es - ru - ko - fr - ja - pt - tr - pl - ca - nl - ar - sv - it - id - hi - fi - vi - he - uk - el - ms - cs - ro - da - hu - ta - false - th - ur - hr - bg - lt - la - mi - ml - cy - sk - te - fa - lv - bn - sr - az - sl - kn - et - mk - br - eu - is - hy - ne - mn - bs - kk - sq - sw - gl - mr - pa - si - km - sn - yo - so - af - oc - ka - be - tg - sd - gu - am - yi - lo - uz - fo - ht - ps - tk - nn - mt - sa - lb - my - bo - tl - mg - as - tt - haw - ln - ha - ba - jw - su license: apache-2.0 pipeline_tag: automatic-speech-recognition tags: - audio - automatic-speech-recognition - hf-asr-leaderboard widget: - example_title: Librispeech sample 1 src: https://cdn-media.huggingface.co/speech_samples/sample1.flac - example_title: Librispeech sample 2 src: https://cdn-media.huggingface.co/speech_samples/sample2.flac model-index: - name: whisper-base results: - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (clean) type: librispeech_asr config: clean split: test args: language: en metrics: - type: wer value: 5.008769117619326 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (other) type: librispeech_asr config: other split: test args: language: en metrics: - type: wer value: 12.84936273212057 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: Common Voice 11.0 type: mozilla-foundation/common_voice_11_0 config: hi split: test args: language: hi metrics: - type: wer value: 131 name: Test WER --- # Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). Disclaimer: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the same language as the audio. For speech translation, the model predicts transcriptions to a different language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: \| Size \| Parameters \| English-only \| Multilingual \| \|----------\|------------\|------------------------------------------------------\|-----------------------------------------------------\| \| tiny \| 39 M \| [✓](https://huggingface.co/openai/whisper-tiny.en) \| [✓](https://huggingface.co/openai/whisper-tiny) \| \| base \| 74 M \| [✓](https://huggingface.co/openai/whisper-base.en) \| [✓](https://huggingface.co/openai/whisper-base) \| \| small \| 244 M \| [✓](https://huggingface.co/openai/whisper-small.en) \| [✓](https://huggingface.co/openai/whisper-small) \| \| medium \| 769 M \| [✓](https://huggingface.co/openai/whisper-medium.en) \| [✓](https://huggingface.co/openai/whisper-medium) \| \| large \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large) \| \| large-v2 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v2) \| # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<\|startoftranscript\|>` token 2. The second token is the language token (e.g. `<\|en\|>` for English) 3. The third token is the "task token". It can take one of two values: `<\|transcribe\|>` for speech recognition or `<\|translate\|>` for speech translation 4. In addition, a `<\|notimestamps\|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <\|startoftranscript\|> <\|en\|> <\|transcribe\|> <\|notimestamps\|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<\|startoftranscript\|><\|en\|><\|transcribe\|><\|notimestamps\|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<\|startoftranscript\|><\|fr\|><\|transcribe\|><\|notimestamps\|> Un vrai travail intéressant va enfin être mené sur ce sujet.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through fine-tuning. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```	null	Non_BioNLP	# Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). Disclaimer: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the same language as the audio. For speech translation, the model predicts transcriptions to a different language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: \| Size \| Parameters \| English-only \| Multilingual \| \|----------\|------------\|------------------------------------------------------\|-----------------------------------------------------\| \| tiny \| 39 M \| [✓](https://huggingface.co/openai/whisper-tiny.en) \| [✓](https://huggingface.co/openai/whisper-tiny) \| \| base \| 74 M \| [✓](https://huggingface.co/openai/whisper-base.en) \| [✓](https://huggingface.co/openai/whisper-base) \| \| small \| 244 M \| [✓](https://huggingface.co/openai/whisper-small.en) \| [✓](https://huggingface.co/openai/whisper-small) \| \| medium \| 769 M \| [✓](https://huggingface.co/openai/whisper-medium.en) \| [✓](https://huggingface.co/openai/whisper-medium) \| \| large \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large) \| \| large-v2 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v2) \| # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<\|startoftranscript\|>` token 2. The second token is the language token (e.g. `<\|en\|>` for English) 3. The third token is the "task token". It can take one of two values: `<\|transcribe\|>` for speech recognition or `<\|translate\|>` for speech translation 4. In addition, a `<\|notimestamps\|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <\|startoftranscript\|> <\|en\|> <\|transcribe\|> <\|notimestamps\|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<\|startoftranscript\|><\|en\|><\|transcribe\|><\|notimestamps\|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<\|startoftranscript\|><\|fr\|><\|transcribe\|><\|notimestamps\|> Un vrai travail intéressant va enfin être mené sur ce sujet.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through fine-tuning. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```	{"language": ["en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", false, "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su"], "license": "apache-2.0", "pipeline_tag": "automatic-speech-recognition", "tags": ["audio", "automatic-speech-recognition", "hf-asr-leaderboard"], "widget": [{"example_title": "Librispeech sample 1", "src": "https://cdn-media.huggingface.co/speech_samples/sample1.flac"}, {"example_title": "Librispeech sample 2", "src": "https://cdn-media.huggingface.co/speech_samples/sample2.flac"}], "model-index": [{"name": "whisper-base", "results": [{"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (clean)", "type": "librispeech_asr", "config": "clean", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 5.008769117619326, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (other)", "type": "librispeech_asr", "config": "other", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 12.84936273212057, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "Common Voice 11.0", "type": "mozilla-foundation/common_voice_11_0", "config": "hi", "split": "test", "args": {"language": "hi"}}, "metrics": [{"type": "wer", "value": 131, "name": "Test WER"}]}]}]}	task	[ "TRANSLATION" ]	44,562
yashugupta786/bart_large_xsum_samsum_conv_summarizer	yashugupta786	summarization	[ "transformers", "pytorch", "bart", "text2text-generation", "summarization", "conversational", "seq2seq", "bart large", "en", "dataset:samsum", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-04-05T08:22:21Z	2023-04-06T12:49:32+00:00	119	0	--- datasets: - samsum language: - en library_name: transformers metrics: - rouge pipeline_tag: summarization tags: - summarization - conversational - seq2seq - bart large widget: - text: 'Hannah: Hey, do you have Betty''s number? Amanda: Lemme check Amanda: Sorry, can''t find it. Amanda: Ask Larry Amanda: He called her last time we were at the park together Hannah: I don''t know him well Amanda: Don''t be shy, he''s very nice Hannah: If you say so.. Hannah: I''d rather you texted him Amanda: Just text him 🙂 Hannah: Urgh.. Alright Hannah: Bye Amanda: Bye bye ' model-index: - name: bart-large-xsum-samsum-conversational_summarizer results: - task: type: abstractive-text-summarization name: Abstractive Text Summarization dataset: name: 'SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization' type: samsum metrics: - type: rouge-1 value: 54.3921 name: Validation ROUGE-1 - type: rouge-2 value: 29.8078 name: Validation ROUGE-2 - type: rouge-l value: 45.1543 name: Validation ROUGE-L - type: rouge-1 value: 53.3059 name: Test ROUGE-1 - type: rouge-2 value: 28.355 name: Test ROUGE-2 - type: rouge-l value: 44.0953 name: Test ROUGE-L --- ## Usage ```python from transformers import pipeline summarizer_pipe = pipeline("summarization", model="yashugupta786/bart_large_xsum_samsum_conv_summarizer") conversation_data = '''Hannah: Hey, do you have Betty's number? Amanda: Lemme check Amanda: Sorry, can't find it. Amanda: Ask Larry Amanda: He called her last time we were at the park together Hannah: I don't know him well Amanda: Don't be shy, he's very nice Hannah: If you say so.. Hannah: I'd rather you texted him Amanda: Just text him 🙂 Hannah: Urgh.. Alright Hannah: Bye Amanda: Bye bye ''' summarizer_pipe(conversation_data) ``` ## Results \| key \| value \| \| --- \| ----- \| \| eval_rouge1 \| 54.3921 \| \| eval_rouge2 \| 29.8078 \| \| eval_rougeL \| 45.1543 \| \| eval_rougeLsum \| 49.942 \| \| test_rouge1 \| 53.3059 \| \| test_rouge2 \| 28.355 \| \| test_rougeL \| 44.0953 \| \| test_rougeLsum \| 48.9246 \| All the metric Rouge1,2,L are computed using precison and recall then computed the F measure for these Rouge recall= no of overlaping words/total no of referenced humman annotated words Rouge precision= no of overlaping words/total no of candidate machine predicted words	null	Non_BioNLP	## Usage ```python from transformers import pipeline summarizer_pipe = pipeline("summarization", model="yashugupta786/bart_large_xsum_samsum_conv_summarizer") conversation_data = '''Hannah: Hey, do you have Betty's number? Amanda: Lemme check Amanda: Sorry, can't find it. Amanda: Ask Larry Amanda: He called her last time we were at the park together Hannah: I don't know him well Amanda: Don't be shy, he's very nice Hannah: If you say so.. Hannah: I'd rather you texted him Amanda: Just text him 🙂 Hannah: Urgh.. Alright Hannah: Bye Amanda: Bye bye ''' summarizer_pipe(conversation_data) ``` ## Results \| key \| value \| \| --- \| ----- \| \| eval_rouge1 \| 54.3921 \| \| eval_rouge2 \| 29.8078 \| \| eval_rougeL \| 45.1543 \| \| eval_rougeLsum \| 49.942 \| \| test_rouge1 \| 53.3059 \| \| test_rouge2 \| 28.355 \| \| test_rougeL \| 44.0953 \| \| test_rougeLsum \| 48.9246 \| All the metric Rouge1,2,L are computed using precison and recall then computed the F measure for these Rouge recall= no of overlaping words/total no of referenced humman annotated words Rouge precision= no of overlaping words/total no of candidate machine predicted words	{"datasets": ["samsum"], "language": ["en"], "library_name": "transformers", "metrics": ["rouge"], "pipeline_tag": "summarization", "tags": ["summarization", "conversational", "seq2seq", "bart large"], "widget": [{"text": "Hannah: Hey, do you have Betty's number?\nAmanda: Lemme check\nAmanda: Sorry, can't find it.\nAmanda: Ask Larry\nAmanda: He called her last time we were at the park together\nHannah: I don't know him well\nAmanda: Don't be shy, he's very nice\nHannah: If you say so..\nHannah: I'd rather you texted him\nAmanda: Just text him 🙂\nHannah: Urgh.. Alright\nHannah: Bye\nAmanda: Bye bye\n"}], "model-index": [{"name": "bart-large-xsum-samsum-conversational_summarizer", "results": [{"task": {"type": "abstractive-text-summarization", "name": "Abstractive Text Summarization"}, "dataset": {"name": "SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization", "type": "samsum"}, "metrics": [{"type": "rouge-1", "value": 54.3921, "name": "Validation ROUGE-1"}, {"type": "rouge-2", "value": 29.8078, "name": "Validation ROUGE-2"}, {"type": "rouge-l", "value": 45.1543, "name": "Validation ROUGE-L"}, {"type": "rouge-1", "value": 53.3059, "name": "Test ROUGE-1"}, {"type": "rouge-2", "value": 28.355, "name": "Test ROUGE-2"}, {"type": "rouge-l", "value": 44.0953, "name": "Test ROUGE-L"}]}]}]}	task	[ "SUMMARIZATION" ]	44,563
tunaozates/bert-base-uncased-finetuned-cola	tunaozates	text-classification	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-05-05T18:25:08Z	2023-05-05T18:34:57+00:00	8	0	--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: bert-base-uncased-finetuned-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: validation args: cola metrics: - type: matthews_correlation value: 0.5372712841497043 name: Matthews Correlation --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4558 - Matthews Correlation: 0.5373 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.4965 \| 1.0 \| 535 \| 0.4558 \| 0.5373 \| ### Framework versions - Transformers 4.28.1 - Pytorch 1.13.1+cu117 - Datasets 2.11.0 - Tokenizers 0.12.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4558 - Matthews Correlation: 0.5373 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.4965 \| 1.0 \| 535 \| 0.4558 \| 0.5373 \| ### Framework versions - Transformers 4.28.1 - Pytorch 1.13.1+cu117 - Datasets 2.11.0 - Tokenizers 0.12.1	{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-uncased-finetuned-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "validation", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.5372712841497043, "name": "Matthews Correlation"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,564
prithivMLmods/Bellatrix-Tiny-1B	prithivMLmods	text-generation	[ "transformers", "safetensors", "llama", "text-generation", "reason", "tiny", "llama3.2", "conversational", "en", "base_model:meta-llama/Llama-3.2-1B-Instruct", "base_model:finetune:meta-llama/Llama-3.2-1B-Instruct", "license:llama3.2", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2025-01-26T05:27:07Z	2025-01-27T13:55:42+00:00	267	2	--- base_model: - meta-llama/Llama-3.2-1B-Instruct language: - en library_name: transformers license: llama3.2 pipeline_tag: text-generation tags: - reason - tiny - llama - llama3.2 --- ![logo.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/Rqm-Qx8AvbHFFbFbVY93X.png) <pre align="center"> ____ ____ __ __ __ ____ ____ ____ _ _ ( _ \( ___)( ) ( ) /__\ (_ _)( _ \(_ _)( \/ ) ) _ < )__) )(__ )(__ /(__)\ )( ) / _)(_ ) ( (____/(____)(____)(____)(__)(__)(__) (_)\_)(____)(_/\_) </pre> # Bellatrix-Tiny-1B Bellatrix is based on a reasoning-based model designed for the QWQ synthetic dataset entries. The pipeline's instruction-tuned, text-only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. These models outperform many of the available open-source options. Bellatrix is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions utilize supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). # Use with transformers Starting with `transformers >= 4.43.0` onward, you can run conversational inference using the Transformers `pipeline` abstraction or by leveraging the Auto classes with the `generate()` function. Make sure to update your transformers installation via `pip install --upgrade transformers`. ```python import torch from transformers import pipeline model_id = "prithivMLmods/Bellatrix-Tiny-1B" pipe = pipeline( "text-generation", model=model_id, torch_dtype=torch.bfloat16, device_map="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] outputs = pipe( messages, max_new_tokens=256, ) print(outputs[0]["generated_text"][-1]) ``` Note: You can also find detailed recipes on how to use the model locally, with `torch.compile()`, assisted generations, quantised and more at [`huggingface-llama-recipes`](https://github.com/huggingface/huggingface-llama-recipes) # Intended Use Bellatrix is designed for applications that require advanced reasoning and multilingual dialogue capabilities. It is particularly suitable for: - Agentic Retrieval: Enabling intelligent retrieval of relevant information in a dialogue or query-response system. - Summarization Tasks: Condensing large bodies of text into concise summaries for easier comprehension. - Multilingual Use Cases: Supporting conversations in multiple languages with high accuracy and coherence. - Instruction-Based Applications: Following complex, context-aware instructions to generate precise outputs in a variety of scenarios. # Limitations Despite its capabilities, Bellatrix has some limitations: 1. Domain Specificity: While it performs well on general tasks, its performance may degrade with highly specialized or niche datasets. 2. Dependence on Training Data: It is only as good as the quality and diversity of its training data, which may lead to biases or inaccuracies. 3. Computational Resources: The model’s optimized transformer architecture can be resource-intensive, requiring significant computational power for fine-tuning and inference. 4. Language Coverage: While multilingual, some languages or dialects may have limited support or lower performance compared to widely used ones. 5. Real-World Contexts: It may struggle with understanding nuanced or ambiguous real-world scenarios not covered during training.	null	Non_BioNLP	![logo.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/Rqm-Qx8AvbHFFbFbVY93X.png) <pre align="center"> ____ ____ __ __ __ ____ ____ ____ _ _ ( _ \( ___)( ) ( ) /__\ (_ _)( _ \(_ _)( \/ ) ) _ < )__) )(__ )(__ /(__)\ )( ) / _)(_ ) ( (____/(____)(____)(____)(__)(__)(__) (_)\_)(____)(_/\_) </pre> # Bellatrix-Tiny-1B Bellatrix is based on a reasoning-based model designed for the QWQ synthetic dataset entries. The pipeline's instruction-tuned, text-only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. These models outperform many of the available open-source options. Bellatrix is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions utilize supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). # Use with transformers Starting with `transformers >= 4.43.0` onward, you can run conversational inference using the Transformers `pipeline` abstraction or by leveraging the Auto classes with the `generate()` function. Make sure to update your transformers installation via `pip install --upgrade transformers`. ```python import torch from transformers import pipeline model_id = "prithivMLmods/Bellatrix-Tiny-1B" pipe = pipeline( "text-generation", model=model_id, torch_dtype=torch.bfloat16, device_map="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] outputs = pipe( messages, max_new_tokens=256, ) print(outputs[0]["generated_text"][-1]) ``` Note: You can also find detailed recipes on how to use the model locally, with `torch.compile()`, assisted generations, quantised and more at [`huggingface-llama-recipes`](https://github.com/huggingface/huggingface-llama-recipes) # Intended Use Bellatrix is designed for applications that require advanced reasoning and multilingual dialogue capabilities. It is particularly suitable for: - Agentic Retrieval: Enabling intelligent retrieval of relevant information in a dialogue or query-response system. - Summarization Tasks: Condensing large bodies of text into concise summaries for easier comprehension. - Multilingual Use Cases: Supporting conversations in multiple languages with high accuracy and coherence. - Instruction-Based Applications: Following complex, context-aware instructions to generate precise outputs in a variety of scenarios. # Limitations Despite its capabilities, Bellatrix has some limitations: 1. Domain Specificity: While it performs well on general tasks, its performance may degrade with highly specialized or niche datasets. 2. Dependence on Training Data: It is only as good as the quality and diversity of its training data, which may lead to biases or inaccuracies. 3. Computational Resources: The model’s optimized transformer architecture can be resource-intensive, requiring significant computational power for fine-tuning and inference. 4. Language Coverage: While multilingual, some languages or dialects may have limited support or lower performance compared to widely used ones. 5. Real-World Contexts: It may struggle with understanding nuanced or ambiguous real-world scenarios not covered during training.	{"base_model": ["meta-llama/Llama-3.2-1B-Instruct"], "language": ["en"], "library_name": "transformers", "license": "llama3.2", "pipeline_tag": "text-generation", "tags": ["reason", "tiny", "llama", "llama3.2"]}	task	[ "SUMMARIZATION" ]	44,565
Terps/mt5-small-finetuned-amazon-en-es	Terps	summarization	[ "transformers", "pytorch", "mt5", "text2text-generation", "summarization", "generated_from_trainer", "base_model:google/mt5-small", "base_model:finetune:google/mt5-small", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-09-07T18:42:36Z	2023-09-07T19:46:31+00:00	15	0	--- base_model: google/mt5-small license: apache-2.0 metrics: - rouge tags: - summarization - generated_from_trainer model-index: - name: mt5-small-finetuned-amazon-en-es results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # mt5-small-finetuned-amazon-en-es This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.0279 - Rouge1: 16.4284 - Rouge2: 7.8601 - Rougel: 16.0029 - Rougelsum: 16.0246 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5.6e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Rouge1 \| Rouge2 \| Rougel \| Rougelsum \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:-------:\|:------:\|:-------:\|:---------:\| \| 4.4194 \| 1.0 \| 1209 \| 3.3097 \| 14.9867 \| 6.4886 \| 14.4174 \| 14.4646 \| \| 3.8132 \| 2.0 \| 2418 \| 3.1602 \| 16.1474 \| 7.9815 \| 15.5342 \| 15.6445 \| \| 3.5412 \| 3.0 \| 3627 \| 3.0789 \| 17.4468 \| 8.8014 \| 16.9142 \| 17.002 \| \| 3.3861 \| 4.0 \| 4836 \| 3.0775 \| 15.903 \| 7.4423 \| 15.4008 \| 15.3871 \| \| 3.2952 \| 5.0 \| 6045 \| 3.0480 \| 15.8646 \| 7.3936 \| 15.3989 \| 15.4395 \| \| 3.2155 \| 6.0 \| 7254 \| 3.0354 \| 16.5887 \| 8.0624 \| 16.2377 \| 16.2562 \| \| 3.1896 \| 7.0 \| 8463 \| 3.0273 \| 17.1092 \| 8.5391 \| 16.6507 \| 16.7272 \| \| 3.1594 \| 8.0 \| 9672 \| 3.0279 \| 16.4284 \| 7.8601 \| 16.0029 \| 16.0246 \| ### Framework versions - Transformers 4.33.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.5 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # mt5-small-finetuned-amazon-en-es This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.0279 - Rouge1: 16.4284 - Rouge2: 7.8601 - Rougel: 16.0029 - Rougelsum: 16.0246 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5.6e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Rouge1 \| Rouge2 \| Rougel \| Rougelsum \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:-------:\|:------:\|:-------:\|:---------:\| \| 4.4194 \| 1.0 \| 1209 \| 3.3097 \| 14.9867 \| 6.4886 \| 14.4174 \| 14.4646 \| \| 3.8132 \| 2.0 \| 2418 \| 3.1602 \| 16.1474 \| 7.9815 \| 15.5342 \| 15.6445 \| \| 3.5412 \| 3.0 \| 3627 \| 3.0789 \| 17.4468 \| 8.8014 \| 16.9142 \| 17.002 \| \| 3.3861 \| 4.0 \| 4836 \| 3.0775 \| 15.903 \| 7.4423 \| 15.4008 \| 15.3871 \| \| 3.2952 \| 5.0 \| 6045 \| 3.0480 \| 15.8646 \| 7.3936 \| 15.3989 \| 15.4395 \| \| 3.2155 \| 6.0 \| 7254 \| 3.0354 \| 16.5887 \| 8.0624 \| 16.2377 \| 16.2562 \| \| 3.1896 \| 7.0 \| 8463 \| 3.0273 \| 17.1092 \| 8.5391 \| 16.6507 \| 16.7272 \| \| 3.1594 \| 8.0 \| 9672 \| 3.0279 \| 16.4284 \| 7.8601 \| 16.0029 \| 16.0246 \| ### Framework versions - Transformers 4.33.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.5 - Tokenizers 0.13.3	{"base_model": "google/mt5-small", "license": "apache-2.0", "metrics": ["rouge"], "tags": ["summarization", "generated_from_trainer"], "model-index": [{"name": "mt5-small-finetuned-amazon-en-es", "results": []}]}	task	[ "SUMMARIZATION" ]	44,566
PygTesting/sum_qlora_4pochs	PygTesting	null	[ "peft", "mistral", "generated_from_trainer", "base_model:mistralai/Mistral-Nemo-Base-2407", "base_model:adapter:mistralai/Mistral-Nemo-Base-2407", "license:apache-2.0", "4-bit", "bitsandbytes", "region:us" ]	2024-08-27T19:23:52Z	2024-08-27T19:24:05+00:00	3	0	--- base_model: mistralai/Mistral-Nemo-Base-2407 library_name: peft license: apache-2.0 tags: - generated_from_trainer model-index: - name: qlora_outputs results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml base_model: mistralai/Mistral-Nemo-Base-2407 model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer load_in_8bit: false load_in_4bit: true strict: false plugins: - axolotl.integrations.liger.LigerPlugin liger_rope: true liger_rms_norm: true liger_swiglu: true liger_fused_linear_cross_entropy: true adapter: qlora lora_r: 32 lora_alpha: 16 lora_dropout: 0.05 lora_target_linear: true lora_fan_in_fan_out: datasets: - path: /home/austin/disk1/summaries_fixed.jsonl type: sharegpt dataset_prepared_path: last_run_prepared val_set_size: 0.01 output_dir: ./qlora_outputs sequence_len: 8192 sample_packing: true eval_sample_packing: false pad_to_sequence_len: true wandb_project: summarization-qlora wandb_entity: wandb_watch: wandb_name: actual_run1 wandb_log_model: #unsloth_cross_entropy_loss: true gradient_accumulation_steps: 1 micro_batch_size: 1 num_epochs: 4 optimizer: adamw_bnb_8bit lr_scheduler: cosine learning_rate: 0.0002 train_on_inputs: false group_by_length: false bf16: auto fp16: tf32: false gradient_checkpointing: true gradient_checkpointing_kwargs: use_reentrant: false early_stopping_patience: resume_from_checkpoint: logging_steps: 1 xformers_attention: false flash_attention: true loss_watchdog_threshold: 5.0 loss_watchdog_patience: 3 warmup_steps: 25 evals_per_epoch: 4 eval_table_size: saves_per_epoch: 4 debug: deepspeed: ./deepspeed_configs/zero2.json weight_decay: 0.0 fsdp: # - full_shard # - auto_wrap fsdp_config: # fsdp_limit_all_gathers: true # fsdp_activation_checkpointing: true # fsdp_sync_module_states: true # fsdp_offload_params: false # fsdp_use_orig_params: false # fsdp_cpu_ram_efficient_loading: false # fsdp_transformer_layer_cls_to_wrap: MistralDecoderLayer # fsdp_state_dict_type: FULL_STATE_DICT # fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP special_tokens: pad_token: </s> ``` </details><br> # qlora_outputs This model is a fine-tuned version of [mistralai/Mistral-Nemo-Base-2407](https://huggingface.co/mistralai/Mistral-Nemo-Base-2407) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.5617 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - distributed_type: multi-GPU - num_devices: 4 - total_train_batch_size: 4 - total_eval_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 25 - num_epochs: 4 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:------:\|:----:\|:---------------:\| \| 2.0177 \| 0.0014 \| 1 \| 1.6514 \| \| 1.6259 \| 0.2507 \| 177 \| 1.2032 \| \| 1.4232 \| 0.5014 \| 354 \| 1.1897 \| \| 1.6835 \| 0.7521 \| 531 \| 1.1985 \| \| 1.6514 \| 1.0028 \| 708 \| 1.1874 \| \| 1.4538 \| 1.2365 \| 885 \| 1.2166 \| \| 1.2421 \| 1.4873 \| 1062 \| 1.2224 \| \| 1.2844 \| 1.7380 \| 1239 \| 1.2330 \| \| 1.4152 \| 1.9887 \| 1416 \| 1.2345 \| \| 1.1668 \| 2.2252 \| 1593 \| 1.3476 \| \| 1.1249 \| 2.4759 \| 1770 \| 1.3608 \| \| 0.921 \| 2.7266 \| 1947 \| 1.3793 \| \| 0.7824 \| 2.9773 \| 2124 \| 1.3906 \| \| 1.1759 \| 3.2040 \| 2301 \| 1.5438 \| \| 0.6625 \| 3.4547 \| 2478 \| 1.5644 \| \| 0.8959 \| 3.7054 \| 2655 \| 1.5617 \| ### Framework versions - PEFT 0.12.0 - Transformers 4.44.0 - Pytorch 2.3.0+cu121 - Datasets 2.20.0 - Tokenizers 0.19.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml base_model: mistralai/Mistral-Nemo-Base-2407 model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer load_in_8bit: false load_in_4bit: true strict: false plugins: - axolotl.integrations.liger.LigerPlugin liger_rope: true liger_rms_norm: true liger_swiglu: true liger_fused_linear_cross_entropy: true adapter: qlora lora_r: 32 lora_alpha: 16 lora_dropout: 0.05 lora_target_linear: true lora_fan_in_fan_out: datasets: - path: /home/austin/disk1/summaries_fixed.jsonl type: sharegpt dataset_prepared_path: last_run_prepared val_set_size: 0.01 output_dir: ./qlora_outputs sequence_len: 8192 sample_packing: true eval_sample_packing: false pad_to_sequence_len: true wandb_project: summarization-qlora wandb_entity: wandb_watch: wandb_name: actual_run1 wandb_log_model: #unsloth_cross_entropy_loss: true gradient_accumulation_steps: 1 micro_batch_size: 1 num_epochs: 4 optimizer: adamw_bnb_8bit lr_scheduler: cosine learning_rate: 0.0002 train_on_inputs: false group_by_length: false bf16: auto fp16: tf32: false gradient_checkpointing: true gradient_checkpointing_kwargs: use_reentrant: false early_stopping_patience: resume_from_checkpoint: logging_steps: 1 xformers_attention: false flash_attention: true loss_watchdog_threshold: 5.0 loss_watchdog_patience: 3 warmup_steps: 25 evals_per_epoch: 4 eval_table_size: saves_per_epoch: 4 debug: deepspeed: ./deepspeed_configs/zero2.json weight_decay: 0.0 fsdp: # - full_shard # - auto_wrap fsdp_config: # fsdp_limit_all_gathers: true # fsdp_activation_checkpointing: true # fsdp_sync_module_states: true # fsdp_offload_params: false # fsdp_use_orig_params: false # fsdp_cpu_ram_efficient_loading: false # fsdp_transformer_layer_cls_to_wrap: MistralDecoderLayer # fsdp_state_dict_type: FULL_STATE_DICT # fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP special_tokens: pad_token: </s> ``` </details><br> # qlora_outputs This model is a fine-tuned version of [mistralai/Mistral-Nemo-Base-2407](https://huggingface.co/mistralai/Mistral-Nemo-Base-2407) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.5617 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - distributed_type: multi-GPU - num_devices: 4 - total_train_batch_size: 4 - total_eval_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 25 - num_epochs: 4 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:------:\|:----:\|:---------------:\| \| 2.0177 \| 0.0014 \| 1 \| 1.6514 \| \| 1.6259 \| 0.2507 \| 177 \| 1.2032 \| \| 1.4232 \| 0.5014 \| 354 \| 1.1897 \| \| 1.6835 \| 0.7521 \| 531 \| 1.1985 \| \| 1.6514 \| 1.0028 \| 708 \| 1.1874 \| \| 1.4538 \| 1.2365 \| 885 \| 1.2166 \| \| 1.2421 \| 1.4873 \| 1062 \| 1.2224 \| \| 1.2844 \| 1.7380 \| 1239 \| 1.2330 \| \| 1.4152 \| 1.9887 \| 1416 \| 1.2345 \| \| 1.1668 \| 2.2252 \| 1593 \| 1.3476 \| \| 1.1249 \| 2.4759 \| 1770 \| 1.3608 \| \| 0.921 \| 2.7266 \| 1947 \| 1.3793 \| \| 0.7824 \| 2.9773 \| 2124 \| 1.3906 \| \| 1.1759 \| 3.2040 \| 2301 \| 1.5438 \| \| 0.6625 \| 3.4547 \| 2478 \| 1.5644 \| \| 0.8959 \| 3.7054 \| 2655 \| 1.5617 \| ### Framework versions - PEFT 0.12.0 - Transformers 4.44.0 - Pytorch 2.3.0+cu121 - Datasets 2.20.0 - Tokenizers 0.19.1	{"base_model": "mistralai/Mistral-Nemo-Base-2407", "library_name": "peft", "license": "apache-2.0", "tags": ["generated_from_trainer"], "model-index": [{"name": "qlora_outputs", "results": []}]}	task	[ "SUMMARIZATION" ]	44,567
ilkekas/bert-base-uncased-finetuned2-cola	ilkekas	text-classification	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-05-06T09:26:03Z	2023-05-06T10:31:07+00:00	9	0	--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: bert-base-uncased-finetuned2-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: validation args: cola metrics: - type: matthews_correlation value: 0.5650459791482846 name: Matthews Correlation --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-uncased-finetuned2-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5176 - Matthews Correlation: 0.5650 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2.6781109393881056e-06 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.5726 \| 1.0 \| 535 \| 0.5090 \| 0.3912 \| \| 0.4467 \| 2.0 \| 1070 \| 0.4536 \| 0.5024 \| \| 0.3891 \| 3.0 \| 1605 \| 0.5093 \| 0.4943 \| \| 0.3387 \| 4.0 \| 2140 \| 0.4927 \| 0.5365 \| \| 0.3177 \| 5.0 \| 2675 \| 0.4897 \| 0.5624 \| \| 0.2853 \| 6.0 \| 3210 \| 0.5176 \| 0.5650 \| \| 0.2718 \| 7.0 \| 3745 \| 0.5440 \| 0.5524 \| \| 0.2532 \| 8.0 \| 4280 \| 0.5431 \| 0.5602 \| ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-uncased-finetuned2-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5176 - Matthews Correlation: 0.5650 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2.6781109393881056e-06 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.5726 \| 1.0 \| 535 \| 0.5090 \| 0.3912 \| \| 0.4467 \| 2.0 \| 1070 \| 0.4536 \| 0.5024 \| \| 0.3891 \| 3.0 \| 1605 \| 0.5093 \| 0.4943 \| \| 0.3387 \| 4.0 \| 2140 \| 0.4927 \| 0.5365 \| \| 0.3177 \| 5.0 \| 2675 \| 0.4897 \| 0.5624 \| \| 0.2853 \| 6.0 \| 3210 \| 0.5176 \| 0.5650 \| \| 0.2718 \| 7.0 \| 3745 \| 0.5440 \| 0.5524 \| \| 0.2532 \| 8.0 \| 4280 \| 0.5431 \| 0.5602 \| ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3	{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-uncased-finetuned2-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "validation", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.5650459791482846, "name": "Matthews Correlation"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,568
BojanSimoski/distilbert-base-uncased-finetuned-cola	BojanSimoski	text-classification	[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-08-17T15:48:29Z	2022-08-17T17:34:52+00:00	10	0	--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: train args: cola metrics: - type: matthews_correlation value: 0.5491398222815213 name: Matthews Correlation --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-cola This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5196 - Matthews Correlation: 0.5491 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.5224 \| 1.0 \| 535 \| 0.5262 \| 0.4063 \| \| 0.351 \| 2.0 \| 1070 \| 0.4991 \| 0.4871 \| \| 0.2369 \| 3.0 \| 1605 \| 0.5196 \| 0.5491 \| \| 0.1756 \| 4.0 \| 2140 \| 0.7817 \| 0.5142 \| \| 0.1268 \| 5.0 \| 2675 \| 0.8089 \| 0.5324 \| ### Framework versions - Transformers 4.21.1 - Pytorch 1.12.1+cu113 - Datasets 2.4.0 - Tokenizers 0.12.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-cola This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5196 - Matthews Correlation: 0.5491 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.5224 \| 1.0 \| 535 \| 0.5262 \| 0.4063 \| \| 0.351 \| 2.0 \| 1070 \| 0.4991 \| 0.4871 \| \| 0.2369 \| 3.0 \| 1605 \| 0.5196 \| 0.5491 \| \| 0.1756 \| 4.0 \| 2140 \| 0.7817 \| 0.5142 \| \| 0.1268 \| 5.0 \| 2675 \| 0.8089 \| 0.5324 \| ### Framework versions - Transformers 4.21.1 - Pytorch 1.12.1+cu113 - Datasets 2.4.0 - Tokenizers 0.12.1	{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "train", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.5491398222815213, "name": "Matthews Correlation"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,569
coverquick/resume-header-classifier	coverquick	text-classification	[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	2023-04-19T07:11:16Z	2023-04-19T07:17:55+00:00	18	0	--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- # resume-header-classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("resume-header-classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	null	Non_BioNLP	# resume-header-classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("resume-header-classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}	task	[ "TEXT_CLASSIFICATION" ]	44,570
lmstudio-community/internlm2-math-plus-20b-GGUF	lmstudio-community	text-generation	[ "gguf", "math", "text-generation", "en", "zh", "base_model:internlm/internlm2-math-plus-20b", "base_model:quantized:internlm/internlm2-math-plus-20b", "license:other", "endpoints_compatible", "region:us", "imatrix", "conversational" ]	2024-05-30T17:36:00Z	2024-05-30T17:52:27+00:00	317	1	--- base_model: internlm/internlm2-math-plus-20b language: - en - zh license: other pipeline_tag: text-generation tags: - math quantized_by: bartowski lm_studio: param_count: 20b use_case: math release_date: 24-05-2024 model_creator: InternLM prompt_template: ChatML system_prompt: none base_model: InternLM original_repo: internlm/internlm2-math-plus-20b --- ## 💫 Community Model> InternLM2 Math Plus 20b by InternLM 👾 [LM Studio](https://lmstudio.ai) Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on [Discord](https://discord.gg/aPQfnNkxGC). Model creator: [InternLM](https://huggingface.co/internlm)<br> Original model: [internlm2-math-plus-20b](https://huggingface.co/internlm/internlm2-math-plus-20b)<br> GGUF quantization: provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b3001](https://github.com/ggerganov/llama.cpp/releases/tag/b3001)<br> ## Model Summary: InternLM2 Math Plus is a series of math proficient models by InternLM, following up on their original series of math models.<br> This series has state of the art bilingual open-sourced math reasoning models at several sizes. This should be used as a solver, prover, verifier, augmentor, with chain of thought reasoning. ## Prompt template: Choose the `ChatML` preset in your LM Studio. Under the hood, the model will see a prompt that's formatted like so: ``` <\|im_start\|>user {prompt}<\|im_end\|> <\|im_start\|>assistant ``` ## Technical Details Math Plus has improved informal math reasoning performance (chain-of-thought and code-intepreter) and formal math reasoning performance (LEAN 4 translation and LEAN 4 theorem proving).<br> InternLM2-Math are continued pretrained from InternLM2-Base with ~100B high quality math-related tokens and SFT with ~2M bilingual math supervised data.<br> More details can be found here: https://github.com/InternLM/InternLM-Math ## Special thanks 🙏 Special thanks to [Georgi Gerganov](https://github.com/ggerganov) and the whole team working on [llama.cpp](https://github.com/ggerganov/llama.cpp/) 🙏 Special thanks to [Kalomaze](https://github.com/kalomaze) and [Dampf](https://github.com/Dampfinchen) for their work on the dataset (linked [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)) that was used for calculating the imatrix for all sizes. ## Disclaimers LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model. You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models. LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.	null	Non_BioNLP	## 💫 Community Model> InternLM2 Math Plus 20b by InternLM 👾 [LM Studio](https://lmstudio.ai) Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on [Discord](https://discord.gg/aPQfnNkxGC). Model creator: [InternLM](https://huggingface.co/internlm)<br> Original model: [internlm2-math-plus-20b](https://huggingface.co/internlm/internlm2-math-plus-20b)<br> GGUF quantization: provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b3001](https://github.com/ggerganov/llama.cpp/releases/tag/b3001)<br> ## Model Summary: InternLM2 Math Plus is a series of math proficient models by InternLM, following up on their original series of math models.<br> This series has state of the art bilingual open-sourced math reasoning models at several sizes. This should be used as a solver, prover, verifier, augmentor, with chain of thought reasoning. ## Prompt template: Choose the `ChatML` preset in your LM Studio. Under the hood, the model will see a prompt that's formatted like so: ``` <\|im_start\|>user {prompt}<\|im_end\|> <\|im_start\|>assistant ``` ## Technical Details Math Plus has improved informal math reasoning performance (chain-of-thought and code-intepreter) and formal math reasoning performance (LEAN 4 translation and LEAN 4 theorem proving).<br> InternLM2-Math are continued pretrained from InternLM2-Base with ~100B high quality math-related tokens and SFT with ~2M bilingual math supervised data.<br> More details can be found here: https://github.com/InternLM/InternLM-Math ## Special thanks 🙏 Special thanks to [Georgi Gerganov](https://github.com/ggerganov) and the whole team working on [llama.cpp](https://github.com/ggerganov/llama.cpp/) 🙏 Special thanks to [Kalomaze](https://github.com/kalomaze) and [Dampf](https://github.com/Dampfinchen) for their work on the dataset (linked [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)) that was used for calculating the imatrix for all sizes. ## Disclaimers LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model. You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models. LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.	{"base_model": "internlm/internlm2-math-plus-20b", "language": ["en", "zh"], "license": "other", "pipeline_tag": "text-generation", "tags": ["math"], "quantized_by": "bartowski", "lm_studio": {"param_count": "20b", "use_case": "math", "release_date": "24-05-2024", "model_creator": "InternLM", "prompt_template": "ChatML", "system_prompt": "none", "base_model": "InternLM", "original_repo": "internlm/internlm2-math-plus-20b"}}	task	[ "TRANSLATION" ]	44,571
dimcall/distilbert-base-uncased-finetuned-emotion	dimcall	text-classification	[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:emotion", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-10-02T14:55:47Z	2023-10-02T15:06:24+00:00	7	0	--- datasets: - emotion license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion args: split metrics: - type: accuracy value: 0.9265 name: Accuracy - type: f1 value: 0.9265776722587405 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2129 - Accuracy: 0.9265 - F1: 0.9266 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 0.8255 \| 1.0 \| 250 \| 0.3174 \| 0.901 \| 0.8980 \| \| 0.2503 \| 2.0 \| 500 \| 0.2129 \| 0.9265 \| 0.9266 \| ### Framework versions - Transformers 4.16.2 - Pytorch 1.11.0+cu115 - Datasets 1.16.1 - Tokenizers 0.14.0	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2129 - Accuracy: 0.9265 - F1: 0.9266 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 0.8255 \| 1.0 \| 250 \| 0.3174 \| 0.901 \| 0.8980 \| \| 0.2503 \| 2.0 \| 500 \| 0.2129 \| 0.9265 \| 0.9266 \| ### Framework versions - Transformers 4.16.2 - Pytorch 1.11.0+cu115 - Datasets 1.16.1 - Tokenizers 0.14.0	{"datasets": ["emotion"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-emotion", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "args": "split"}, "metrics": [{"type": "accuracy", "value": 0.9265, "name": "Accuracy"}, {"type": "f1", "value": 0.9265776722587405, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,572
TheBloke/medicine-LLM-AWQ	TheBloke	text-generation	[ "transformers", "safetensors", "llama", "text-generation", "biology", "medical", "en", "dataset:Open-Orca/OpenOrca", "dataset:GAIR/lima", "dataset:WizardLM/WizardLM_evol_instruct_V2_196k", "dataset:EleutherAI/pile", "arxiv:2309.09530", "base_model:AdaptLLM/medicine-LLM", "base_model:quantized:AdaptLLM/medicine-LLM", "license:other", "autotrain_compatible", "text-generation-inference", "4-bit", "awq", "region:us" ]	2024-01-15T22:38:37Z	2024-01-15T22:54:43+00:00	100	3	--- base_model: AdaptLLM/medicine-LLM datasets: - Open-Orca/OpenOrca - GAIR/lima - WizardLM/WizardLM_evol_instruct_V2_196k - EleutherAI/pile language: - en license: other metrics: - accuracy model_name: Medicine LLM pipeline_tag: text-generation tags: - biology - medical inference: false model_creator: AdaptLLM model_type: llama prompt_template: '### User Input: {prompt} ### Assistant Output: ' quantized_by: TheBloke --- <!-- markdownlint-disable MD041 --> <!-- header start --> <!-- 200823 --> <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <div style="display: flex; justify-content: space-between; width: 100%;"> <div style="display: flex; flex-direction: column; align-items: flex-start;"> <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p> </div> <div style="display: flex; flex-direction: column; align-items: flex-end;"> <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p> </div> </div> <div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;"> <!-- header end --> # Medicine LLM - AWQ - Model creator: [AdaptLLM](https://huggingface.co/AdaptLLM) - Original model: [Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM) <!-- description start --> ## Description This repo contains AWQ model files for [AdaptLLM's Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM). These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/). ### About AWQ AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings. AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead. It is supported by: - [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ - [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code <!-- description end --> <!-- repositories-available start --> ## Repositories available * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/medicine-LLM-AWQ) * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/medicine-LLM-GPTQ) * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/medicine-LLM-GGUF) * [AdaptLLM's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/AdaptLLM/medicine-LLM) <!-- repositories-available end --> <!-- prompt-template start --> ## Prompt template: AdaptLLM ``` ### User Input: {prompt} ### Assistant Output: ``` <!-- prompt-template end --> <!-- README_AWQ.md-provided-files start --> ## Provided files, and AWQ parameters I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered. Models are released as sharded safetensors files. \| Branch \| Bits \| GS \| AWQ Dataset \| Seq Len \| Size \| \| ------ \| ---- \| -- \| ----------- \| ------- \| ---- \| \| [main](https://huggingface.co/TheBloke/medicine-LLM-AWQ/tree/main) \| 4 \| 128 \| [Medical Medaow WikiDoc](https://huggingface.co/datasets/medalpaca/medical_meadow_wikidoc/viewer/) \| 2048 \| 3.89 GB <!-- README_AWQ.md-provided-files end --> <!-- README_AWQ.md-text-generation-webui start --> ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui) Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui). It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install. 1. Click the Model tab. 2. Under Download custom model or LoRA, enter `TheBloke/medicine-LLM-AWQ`. 3. Click Download. 4. The model will start downloading. Once it's finished it will say "Done". 5. In the top left, click the refresh icon next to Model. 6. In the Model dropdown, choose the model you just downloaded: `medicine-LLM-AWQ` 7. Select Loader: AutoAWQ. 8. Click Load, and the model will load and is now ready for use. 9. If you want any custom settings, set them and then click Save settings for this model followed by Reload the Model in the top right. 10. Once you're ready, click the Text Generation tab and enter a prompt to get started! <!-- README_AWQ.md-text-generation-webui end --> <!-- README_AWQ.md-use-from-vllm start --> ## Multi-user inference server: vLLM Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/). - Please ensure you are using vLLM version 0.2 or later. - When using vLLM as a server, pass the `--quantization awq` parameter. For example: ```shell python3 -m vllm.entrypoints.api_server --model TheBloke/medicine-LLM-AWQ --quantization awq --dtype auto ``` - When using vLLM from Python code, again set `quantization=awq`. For example: ```python from vllm import LLM, SamplingParams prompts = [ "Tell me about AI", "Write a story about llamas", "What is 291 - 150?", "How much wood would a woodchuck chuck if a woodchuck could chuck wood?", ] prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' prompts = [prompt_template.format(prompt=prompt) for prompt in prompts] sampling_params = SamplingParams(temperature=0.8, top_p=0.95) llm = LLM(model="TheBloke/medicine-LLM-AWQ", quantization="awq", dtype="auto") outputs = llm.generate(prompts, sampling_params) # Print the outputs. for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` <!-- README_AWQ.md-use-from-vllm start --> <!-- README_AWQ.md-use-from-tgi start --> ## Multi-user inference server: Hugging Face Text Generation Inference (TGI) Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0` Example Docker parameters: ```shell --model-id TheBloke/medicine-LLM-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096 ``` Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later): ```shell pip3 install huggingface-hub ``` ```python from huggingface_hub import InferenceClient endpoint_url = "https://your-endpoint-url-here" prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' client = InferenceClient(endpoint_url) response = client.text_generation(prompt, max_new_tokens=128, do_sample=True, temperature=0.7, top_p=0.95, top_k=40, repetition_penalty=1.1) print(f"Model output: ", response) ``` <!-- README_AWQ.md-use-from-tgi end --> <!-- README_AWQ.md-use-from-python start --> ## Inference from Python code using Transformers ### Install the necessary packages - Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later. - Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later. ```shell pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0" ``` Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0. If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command: ```shell pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl ``` If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead: ```shell pip3 uninstall -y autoawq git clone https://github.com/casper-hansen/AutoAWQ cd AutoAWQ pip3 install . ``` ### Transformers example code (requires Transformers 4.35.0 and later) ```python from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer model_name_or_path = "TheBloke/medicine-LLM-AWQ" tokenizer = AutoTokenizer.from_pretrained(model_name_or_path) model = AutoModelForCausalLM.from_pretrained( model_name_or_path, low_cpu_mem_usage=True, device_map="cuda:0" ) # Using the text streamer to stream output one token at a time streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True) prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' # Convert prompt to tokens tokens = tokenizer( prompt_template, return_tensors='pt' ).input_ids.cuda() generation_params = { "do_sample": True, "temperature": 0.7, "top_p": 0.95, "top_k": 40, "max_new_tokens": 512, "repetition_penalty": 1.1 } # Generate streamed output, visible one token at a time generation_output = model.generate( tokens, streamer=streamer, generation_params ) # Generation without a streamer, which will include the prompt in the output generation_output = model.generate( tokens, generation_params ) # Get the tokens from the output, decode them, print them token_output = generation_output[0] text_output = tokenizer.decode(token_output) print("model.generate output: ", text_output) # Inference is also possible via Transformers' pipeline from transformers import pipeline pipe = pipeline( "text-generation", model=model, tokenizer=tokenizer, *generation_params ) pipe_output = pipe(prompt_template)[0]['generated_text'] print("pipeline output: ", pipe_output) ``` <!-- README_AWQ.md-use-from-python end --> <!-- README_AWQ.md-compatibility start --> ## Compatibility The files provided are tested to work with: - [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`. - [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later. - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later. - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later. <!-- README_AWQ.md-compatibility end --> <!-- footer start --> <!-- 200823 --> ## Discord For further support, and discussions on these models and AI in general, join us at: [TheBloke AI's Discord server](https://discord.gg/theblokeai) ## Thanks, and how to contribute Thanks to the [chirper.ai](https://chirper.ai) team! Thanks to Clay from [gpus.llm-utils.org](llm-utils)! I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training. If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects. Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits. Patreon: https://patreon.com/TheBlokeAI * Ko-Fi: https://ko-fi.com/TheBlokeAI Special thanks to: Aemon Algiz. Patreon special mentions: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros Thank you to all my generous patrons and donaters! And thank you again to a16z for their generous grant. <!-- footer end --> # Original model card: AdaptLLM's Medicine LLM # Adapt (Large) Language Models to Domains This repo contains the domain-specific base model developed from LLaMA-1-7B, using the method in our paper [Adapting Large Language Models via Reading Comprehension](https://huggingface.co/papers/2309.09530). We explore continued pre-training on domain-specific corpora for large language models. While this approach enriches LLMs with domain knowledge, it significantly hurts their prompting ability for question answering. Inspired by human learning via reading comprehension, we propose a simple method to transform large-scale pre-training corpora into reading comprehension texts, consistently improving prompting performance across tasks in biomedicine, finance, and law domains. Our 7B model competes with much larger domain-specific models like BloombergGPT-50B. ### 🤗 We are currently working hard on developing models across different domains, scales and architectures! Please stay tuned! 🤗 ************************** Updates ************************** * 12/19: Released our [13B base models](https://huggingface.co/AdaptLLM/medicine-LLM-13B) developed from LLaMA-1-13B. * 12/8: Released our [chat models](https://huggingface.co/AdaptLLM/medicine-chat) developed from LLaMA-2-Chat-7B. * 9/18: Released our [paper](https://huggingface.co/papers/2309.09530), [code](https://github.com/microsoft/LMOps), [data](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), and [base models](https://huggingface.co/AdaptLLM/medicine-LLM) developed from LLaMA-1-7B. ## Domain-Specific LLaMA-1 ### LLaMA-1-7B In our paper, we develop three domain-specific models from LLaMA-1-7B, which are also available in Huggingface: [Biomedicine-LLM](https://huggingface.co/AdaptLLM/medicine-LLM), [Finance-LLM](https://huggingface.co/AdaptLLM/finance-LLM) and [Law-LLM](https://huggingface.co/AdaptLLM/law-LLM), the performances of our AdaptLLM compared to other domain-specific LLMs are: <p align='center'> <img src="https://hf.fast360.xyz/production/uploads/650801ced5578ef7e20b33d4/6efPwitFgy-pLTzvccdcP.png" width="700"> </p> ### LLaMA-1-13B Moreover, we scale up our base model to LLaMA-1-13B to see if our method is similarly effective for larger-scale models, and the results are consistently positive too: [Biomedicine-LLM-13B](https://huggingface.co/AdaptLLM/medicine-LLM-13B), [Finance-LLM-13B](https://huggingface.co/AdaptLLM/finance-LLM-13B) and [Law-LLM-13B](https://huggingface.co/AdaptLLM/law-LLM-13B). ## Domain-Specific LLaMA-2-Chat Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our reading comprehension can perfectly fit the data format by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat) For example, to chat with the biomedicine base model (🤗we highly recommend switching to the [chat model](https://huggingface.co/AdaptLLM/medicine-chat) for better response quality!): ```python from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained("AdaptLLM/medicine-LLM") tokenizer = AutoTokenizer.from_pretrained("AdaptLLM/medicine-LLM", use_fast=False) # Put your input here: user_input = '''Question: Which of the following is an example of monosomy? Options: - 46,XX - 47,XXX - 69,XYY - 45,X Please provide your choice first and then provide explanations if possible.''' # Simply use your input as the prompt for base models prompt = user_input inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device) outputs = model.generate(input_ids=inputs, max_length=2048)[0] answer_start = int(inputs.shape[-1]) pred = tokenizer.decode(outputs[answer_start:], skip_special_tokens=True) print(f'### User Input:\n{user_input}\n\n### Assistant Output:\n{pred}') ``` ## Domain-Specific Tasks To easily reproduce our results, we have uploaded the filled-in zero/few-shot input instructions and output completions of each domain-specific task: [biomedicine-tasks](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), [finance-tasks](https://huggingface.co/datasets/AdaptLLM/finance-tasks), and [law-tasks](https://huggingface.co/datasets/AdaptLLM/law-tasks). Note: those filled-in instructions are specifically tailored for models before alignment and do NOT fit for the specific data format required for chat models. ## Citation If you find our work helpful, please cite us: ```bibtex @article{adaptllm, title = {Adapting Large Language Models via Reading Comprehension}, author = {Daixuan Cheng and Shaohan Huang and Furu Wei}, journal = {CoRR}, volume = {abs/2309.09530}, year = {2023} } ```	null	BioNLP	<!-- markdownlint-disable MD041 --> <!-- header start --> <!-- 200823 --> <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <div style="display: flex; justify-content: space-between; width: 100%;"> <div style="display: flex; flex-direction: column; align-items: flex-start;"> <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p> </div> <div style="display: flex; flex-direction: column; align-items: flex-end;"> <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p> </div> </div> <div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;"> <!-- header end --> # Medicine LLM - AWQ - Model creator: [AdaptLLM](https://huggingface.co/AdaptLLM) - Original model: [Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM) <!-- description start --> ## Description This repo contains AWQ model files for [AdaptLLM's Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM). These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/). ### About AWQ AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings. AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead. It is supported by: - [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ - [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code <!-- description end --> <!-- repositories-available start --> ## Repositories available * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/medicine-LLM-AWQ) * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/medicine-LLM-GPTQ) * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/medicine-LLM-GGUF) * [AdaptLLM's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/AdaptLLM/medicine-LLM) <!-- repositories-available end --> <!-- prompt-template start --> ## Prompt template: AdaptLLM ``` ### User Input: {prompt} ### Assistant Output: ``` <!-- prompt-template end --> <!-- README_AWQ.md-provided-files start --> ## Provided files, and AWQ parameters I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered. Models are released as sharded safetensors files. \| Branch \| Bits \| GS \| AWQ Dataset \| Seq Len \| Size \| \| ------ \| ---- \| -- \| ----------- \| ------- \| ---- \| \| [main](https://huggingface.co/TheBloke/medicine-LLM-AWQ/tree/main) \| 4 \| 128 \| [Medical Medaow WikiDoc](https://huggingface.co/datasets/medalpaca/medical_meadow_wikidoc/viewer/) \| 2048 \| 3.89 GB <!-- README_AWQ.md-provided-files end --> <!-- README_AWQ.md-text-generation-webui start --> ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui) Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui). It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install. 1. Click the Model tab. 2. Under Download custom model or LoRA, enter `TheBloke/medicine-LLM-AWQ`. 3. Click Download. 4. The model will start downloading. Once it's finished it will say "Done". 5. In the top left, click the refresh icon next to Model. 6. In the Model dropdown, choose the model you just downloaded: `medicine-LLM-AWQ` 7. Select Loader: AutoAWQ. 8. Click Load, and the model will load and is now ready for use. 9. If you want any custom settings, set them and then click Save settings for this model followed by Reload the Model in the top right. 10. Once you're ready, click the Text Generation tab and enter a prompt to get started! <!-- README_AWQ.md-text-generation-webui end --> <!-- README_AWQ.md-use-from-vllm start --> ## Multi-user inference server: vLLM Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/). - Please ensure you are using vLLM version 0.2 or later. - When using vLLM as a server, pass the `--quantization awq` parameter. For example: ```shell python3 -m vllm.entrypoints.api_server --model TheBloke/medicine-LLM-AWQ --quantization awq --dtype auto ``` - When using vLLM from Python code, again set `quantization=awq`. For example: ```python from vllm import LLM, SamplingParams prompts = [ "Tell me about AI", "Write a story about llamas", "What is 291 - 150?", "How much wood would a woodchuck chuck if a woodchuck could chuck wood?", ] prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' prompts = [prompt_template.format(prompt=prompt) for prompt in prompts] sampling_params = SamplingParams(temperature=0.8, top_p=0.95) llm = LLM(model="TheBloke/medicine-LLM-AWQ", quantization="awq", dtype="auto") outputs = llm.generate(prompts, sampling_params) # Print the outputs. for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` <!-- README_AWQ.md-use-from-vllm start --> <!-- README_AWQ.md-use-from-tgi start --> ## Multi-user inference server: Hugging Face Text Generation Inference (TGI) Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0` Example Docker parameters: ```shell --model-id TheBloke/medicine-LLM-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096 ``` Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later): ```shell pip3 install huggingface-hub ``` ```python from huggingface_hub import InferenceClient endpoint_url = "https://your-endpoint-url-here" prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' client = InferenceClient(endpoint_url) response = client.text_generation(prompt, max_new_tokens=128, do_sample=True, temperature=0.7, top_p=0.95, top_k=40, repetition_penalty=1.1) print(f"Model output: ", response) ``` <!-- README_AWQ.md-use-from-tgi end --> <!-- README_AWQ.md-use-from-python start --> ## Inference from Python code using Transformers ### Install the necessary packages - Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later. - Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later. ```shell pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0" ``` Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0. If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command: ```shell pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl ``` If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead: ```shell pip3 uninstall -y autoawq git clone https://github.com/casper-hansen/AutoAWQ cd AutoAWQ pip3 install . ``` ### Transformers example code (requires Transformers 4.35.0 and later) ```python from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer model_name_or_path = "TheBloke/medicine-LLM-AWQ" tokenizer = AutoTokenizer.from_pretrained(model_name_or_path) model = AutoModelForCausalLM.from_pretrained( model_name_or_path, low_cpu_mem_usage=True, device_map="cuda:0" ) # Using the text streamer to stream output one token at a time streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True) prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' # Convert prompt to tokens tokens = tokenizer( prompt_template, return_tensors='pt' ).input_ids.cuda() generation_params = { "do_sample": True, "temperature": 0.7, "top_p": 0.95, "top_k": 40, "max_new_tokens": 512, "repetition_penalty": 1.1 } # Generate streamed output, visible one token at a time generation_output = model.generate( tokens, streamer=streamer, generation_params ) # Generation without a streamer, which will include the prompt in the output generation_output = model.generate( tokens, generation_params ) # Get the tokens from the output, decode them, print them token_output = generation_output[0] text_output = tokenizer.decode(token_output) print("model.generate output: ", text_output) # Inference is also possible via Transformers' pipeline from transformers import pipeline pipe = pipeline( "text-generation", model=model, tokenizer=tokenizer, *generation_params ) pipe_output = pipe(prompt_template)[0]['generated_text'] print("pipeline output: ", pipe_output) ``` <!-- README_AWQ.md-use-from-python end --> <!-- README_AWQ.md-compatibility start --> ## Compatibility The files provided are tested to work with: - [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`. - [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later. - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later. - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later. <!-- README_AWQ.md-compatibility end --> <!-- footer start --> <!-- 200823 --> ## Discord For further support, and discussions on these models and AI in general, join us at: [TheBloke AI's Discord server](https://discord.gg/theblokeai) ## Thanks, and how to contribute Thanks to the [chirper.ai](https://chirper.ai) team! Thanks to Clay from [gpus.llm-utils.org](llm-utils)! I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training. If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects. Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits. Patreon: https://patreon.com/TheBlokeAI * Ko-Fi: https://ko-fi.com/TheBlokeAI Special thanks to: Aemon Algiz. Patreon special mentions: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros Thank you to all my generous patrons and donaters! And thank you again to a16z for their generous grant. <!-- footer end --> # Original model card: AdaptLLM's Medicine LLM # Adapt (Large) Language Models to Domains This repo contains the domain-specific base model developed from LLaMA-1-7B, using the method in our paper [Adapting Large Language Models via Reading Comprehension](https://huggingface.co/papers/2309.09530). We explore continued pre-training on domain-specific corpora for large language models. While this approach enriches LLMs with domain knowledge, it significantly hurts their prompting ability for question answering. Inspired by human learning via reading comprehension, we propose a simple method to transform large-scale pre-training corpora into reading comprehension texts, consistently improving prompting performance across tasks in biomedicine, finance, and law domains. Our 7B model competes with much larger domain-specific models like BloombergGPT-50B. ### 🤗 We are currently working hard on developing models across different domains, scales and architectures! Please stay tuned! 🤗 ************************** Updates ************************** * 12/19: Released our [13B base models](https://huggingface.co/AdaptLLM/medicine-LLM-13B) developed from LLaMA-1-13B. * 12/8: Released our [chat models](https://huggingface.co/AdaptLLM/medicine-chat) developed from LLaMA-2-Chat-7B. * 9/18: Released our [paper](https://huggingface.co/papers/2309.09530), [code](https://github.com/microsoft/LMOps), [data](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), and [base models](https://huggingface.co/AdaptLLM/medicine-LLM) developed from LLaMA-1-7B. ## Domain-Specific LLaMA-1 ### LLaMA-1-7B In our paper, we develop three domain-specific models from LLaMA-1-7B, which are also available in Huggingface: [Biomedicine-LLM](https://huggingface.co/AdaptLLM/medicine-LLM), [Finance-LLM](https://huggingface.co/AdaptLLM/finance-LLM) and [Law-LLM](https://huggingface.co/AdaptLLM/law-LLM), the performances of our AdaptLLM compared to other domain-specific LLMs are: <p align='center'> <img src="https://hf.fast360.xyz/production/uploads/650801ced5578ef7e20b33d4/6efPwitFgy-pLTzvccdcP.png" width="700"> </p> ### LLaMA-1-13B Moreover, we scale up our base model to LLaMA-1-13B to see if our method is similarly effective for larger-scale models, and the results are consistently positive too: [Biomedicine-LLM-13B](https://huggingface.co/AdaptLLM/medicine-LLM-13B), [Finance-LLM-13B](https://huggingface.co/AdaptLLM/finance-LLM-13B) and [Law-LLM-13B](https://huggingface.co/AdaptLLM/law-LLM-13B). ## Domain-Specific LLaMA-2-Chat Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our reading comprehension can perfectly fit the data format by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat) For example, to chat with the biomedicine base model (🤗we highly recommend switching to the [chat model](https://huggingface.co/AdaptLLM/medicine-chat) for better response quality!): ```python from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained("AdaptLLM/medicine-LLM") tokenizer = AutoTokenizer.from_pretrained("AdaptLLM/medicine-LLM", use_fast=False) # Put your input here: user_input = '''Question: Which of the following is an example of monosomy? Options: - 46,XX - 47,XXX - 69,XYY - 45,X Please provide your choice first and then provide explanations if possible.''' # Simply use your input as the prompt for base models prompt = user_input inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device) outputs = model.generate(input_ids=inputs, max_length=2048)[0] answer_start = int(inputs.shape[-1]) pred = tokenizer.decode(outputs[answer_start:], skip_special_tokens=True) print(f'### User Input:\n{user_input}\n\n### Assistant Output:\n{pred}') ``` ## Domain-Specific Tasks To easily reproduce our results, we have uploaded the filled-in zero/few-shot input instructions and output completions of each domain-specific task: [biomedicine-tasks](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), [finance-tasks](https://huggingface.co/datasets/AdaptLLM/finance-tasks), and [law-tasks](https://huggingface.co/datasets/AdaptLLM/law-tasks). Note: those filled-in instructions are specifically tailored for models before alignment and do NOT fit for the specific data format required for chat models. ## Citation If you find our work helpful, please cite us: ```bibtex @article{adaptllm, title = {Adapting Large Language Models via Reading Comprehension}, author = {Daixuan Cheng and Shaohan Huang and Furu Wei}, journal = {CoRR}, volume = {abs/2309.09530}, year = {2023} } ```	{"base_model": "AdaptLLM/medicine-LLM", "datasets": ["Open-Orca/OpenOrca", "GAIR/lima", "WizardLM/WizardLM_evol_instruct_V2_196k", "EleutherAI/pile"], "language": ["en"], "license": "other", "metrics": ["accuracy"], "model_name": "Medicine LLM", "pipeline_tag": "text-generation", "tags": ["biology", "medical"], "inference": false, "model_creator": "AdaptLLM", "model_type": "llama", "prompt_template": "### User Input:\n{prompt}\n\n### Assistant Output:\n", "quantized_by": "TheBloke"}	task	[ "QUESTION_ANSWERING" ]	44,573
zenml/finetuned-snowflake-arctic-embed-m	zenml	sentence-similarity	[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:1490", "loss:MatryoshkaLoss", "loss:TripletLoss", "en", "arxiv:1908.10084", "arxiv:2205.13147", "arxiv:1703.07737", "base_model:Snowflake/snowflake-arctic-embed-m", "base_model:finetune:Snowflake/snowflake-arctic-embed-m", "license:apache-2.0", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	2024-08-05T19:08:54Z	2024-08-08T14:19:44+00:00	25	1	--- base_model: Snowflake/snowflake-arctic-embed-m datasets: [] language: - en library_name: sentence-transformers license: apache-2.0 metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:1490 - loss:MatryoshkaLoss - loss:TripletLoss widget: - source_sentence: Where is the global configuration directory located in ZenML's default setup? sentences: - '''default'' ... Creating default user ''default'' ...Creating default stack for user ''default'' in workspace default... Active workspace not set. Setting it to the default. The active stack is not set. Setting the active stack to the default workspace stack. Using the default store for the global config. Unable to find ZenML repository in your current working directory (/tmp/folder) or any parent directories. If you want to use an existing repository which is in a different location, set the environment variable ''ZENML_REPOSITORY_PATH''. If you want to create a new repository, run zenml init. Running without an active repository root. Using the default local database. Running with active workspace: ''default'' (global) ┏━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━┓ ┃ ACTIVE │ STACK NAME │ SHARED │ OWNER │ ARTIFACT_STORE │ ORCHESTRATOR ┃ ┠────────┼────────────┼────────┼─────────┼────────────────┼──────────────┨ ┃ 👉 │ default │ ❌ │ default │ default │ default ┃ ┗━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━┛ The following is an example of the layout of the global config directory immediately after initialization: /home/stefan/.config/zenml <- Global Config Directory ├── config.yaml <- Global Configuration Settings └── local_stores <- Every Stack component that stores information \| locally will have its own subdirectory here. ├── a1a0d3d0-d552-4a80-be09-67e5e29be8ee <- e.g. Local Store path for the \| `default` local Artifact Store └── default_zen_store └── zenml.db <- SQLite database where ZenML data (stacks, components, etc) are stored by default. As shown above, the global config directory stores the following information:' - How do you configure the network settings on a Linux server? - 'Reranking for better retrieval Add reranking to your RAG inference for better retrieval performance. Rerankers are a crucial component of retrieval systems that use LLMs. They help improve the quality of the retrieved documents by reordering them based on additional features or scores. In this section, we''ll explore how to add a reranker to your RAG inference pipeline in ZenML. In previous sections, we set up the overall workflow, from data ingestion and preprocessing to embeddings generation and retrieval. We then set up some basic evaluation metrics to assess the performance of our retrieval system. A reranker is a way to squeeze a bit of extra performance out of the system by reordering the retrieved documents based on additional features or scores. As you can see, reranking is an optional addition we make to what we''ve already set up. It''s not strictly necessary, but it can help improve the relevance and quality of the retrieved documents, which in turn can lead to better responses from the LLM. Let''s dive in! PreviousEvaluation in practice NextUnderstanding reranking Last updated 1 month ago' - source_sentence: Where can I find the instructions to enable CUDA for GPU-backed hardware in ZenML SDK Docs? sentences: - 'Migration guide 0.39.1 → 0.41.0 How to migrate your ZenML pipelines and steps from version <=0.39.1 to 0.41.0. ZenML versions 0.40.0 to 0.41.0 introduced a new and more flexible syntax to define ZenML steps and pipelines. This page contains code samples that show you how to upgrade your steps and pipelines to the new syntax. Newer versions of ZenML still work with pipelines and steps defined using the old syntax, but the old syntax is deprecated and will be removed in the future. Overview from typing import Optional from zenml.steps import BaseParameters, Output, StepContext, step from zenml.pipelines import pipeline # Define a Step class MyStepParameters(BaseParameters): param_1: int param_2: Optional[float] = None @step def my_step( params: MyStepParameters, context: StepContext, ) -> Output(int_output=int, str_output=str): result = int(params.param_1 * (params.param_2 or 1)) result_uri = context.get_output_artifact_uri() return result, result_uri # Run the Step separately my_step.entrypoint() # Define a Pipeline @pipeline def my_pipeline(my_step): my_step() step_instance = my_step(params=MyStepParameters(param_1=17)) pipeline_instance = my_pipeline(my_step=step_instance) # Configure and run the Pipeline pipeline_instance.configure(enable_cache=False) schedule = Schedule(...) pipeline_instance.run(schedule=schedule) # Fetch the Pipeline Run last_run = pipeline_instance.get_runs()[0] int_output = last_run.get_step["my_step"].outputs["int_output"].read() from typing import Annotated, Optional, Tuple from zenml import get_step_context, pipeline, step from zenml.client import Client # Define a Step @step def my_step( param_1: int, param_2: Optional[float] = None ) -> Tuple[Annotated[int, "int_output"], Annotated[str, "str_output"]]: result = int(param_1 * (param_2 or 1)) result_uri = get_step_context().get_output_artifact_uri() return result, result_uri # Run the Step separately my_step() # Define a Pipeline @pipeline' - How do I integrate Google Cloud VertexAI into my existing Kubernetes cluster? - ' SDK Docs . Enabling CUDA for GPU-backed hardwareNote that if you wish to use this step operator to run steps on a GPU, you will need to follow the instructions on this page to ensure that it works. It requires adding some extra settings customization and is essential to enable CUDA for the GPU to give its full acceleration. PreviousStep Operators NextGoogle Cloud VertexAI Last updated 19 days ago' - source_sentence: What are the special metadata types supported by ZenML and how are they used? sentences: - 'Special Metadata Types Tracking your metadata. ZenML supports several special metadata types to capture specific kinds of information. Here are examples of how to use the special types Uri, Path, DType, and StorageSize: from zenml.metadata.metadata_types import StorageSize, DType from zenml import log_artifact_metadata log_artifact_metadata( metadata={ "dataset_source": Uri("gs://my-bucket/datasets/source.csv"), "preprocessing_script": Path("/scripts/preprocess.py"), "column_types": { "age": DType("int"), "income": DType("float"), "score": DType("int") }, "processed_data_size": StorageSize(2500000) In this example: Uri is used to indicate a dataset source URI. Path is used to specify the filesystem path to a preprocessing script. DType is used to describe the data types of specific columns. StorageSize is used to indicate the size of the processed data in bytes. These special types help standardize the format of metadata and ensure that it is logged in a consistent and interpretable manner. PreviousGroup metadata NextFetch metadata within steps Last updated 19 days ago' - 'Configure a code repository Connect a Git repository to ZenML to track code changes and collaborate on MLOps projects. Throughout the lifecycle of a MLOps pipeline, it can get quite tiresome to always wait for a Docker build every time after running a pipeline (even if the local Docker cache is used). However, there is a way to just have one pipeline build and keep reusing it until a change to the pipeline environment is made: by connecting a code repository. With ZenML, connecting to a Git repository optimizes the Docker build processes. It also has the added bonus of being a better way of managing repository changes and enabling better code collaboration. Here is how the flow changes when running a pipeline: You trigger a pipeline run on your local machine. ZenML parses the @pipeline function to determine the necessary steps. The local client requests stack information from the ZenML server, which responds with the cloud stack configuration. The local client detects that we''re using a code repository and requests the information from the git repo. Instead of building a new Docker image, the client checks if an existing image can be reused based on the current Git commit hash and other environment metadata. The client initiates a run in the orchestrator, which sets up the execution environment in the cloud, such as a VM. The orchestrator downloads the code directly from the Git repository and uses the existing Docker image to run the pipeline steps. Pipeline steps execute, storing artifacts in the cloud-based artifact store. Throughout the execution, the pipeline run status and metadata are reported back to the ZenML server. By connecting a Git repository, you avoid redundant builds and make your MLOps processes more efficient. Your team can work on the codebase simultaneously, with ZenML handling the version tracking and ensuring that the correct code version is always used for each run. Creating a GitHub Repository' - Can you explain the process of setting up a virtual environment in Python? - source_sentence: What are the benefits of deploying stack components directly from the ZenML CLI? sentences: - '─────────────────────────────────────────────────┨┃ RESOURCE TYPES │ 🔵 gcp-generic, 📦 gcs-bucket, 🌀 kubernetes-cluster, 🐳 docker-registry ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ RESOURCE NAME │ <multiple> ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ SECRET ID │ 4694de65-997b-4929-8831-b49d5e067b97 ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ SESSION DURATION │ N/A ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ EXPIRES IN │ 59m46s ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ OWNER │ default ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ WORKSPACE │ default ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ SHARED │ ➖ ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ CREATED_AT │ 2023-05-19 09:04:33.557126 ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ UPDATED_AT │ 2023-05-19 09:04:33.557127 ┃ ┗━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛ Configuration ┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓' - How do you set up a custom service account for Vertex AI? - '⚒️Manage stacks Deploying your stack components directly from the ZenML CLI The first step in running your pipelines on remote infrastructure is to deploy all the components that you would need, like an MLflow tracking server, a Seldon Core model deployer, and more to your cloud. This can bring plenty of benefits like scalability, reliability, and collaboration. ZenML eases the path to production by providing a seamless way for all tools to interact with others through the use of abstractions. However, one of the most painful parts of this process, from what we see on our Slack and in general, is the deployment of these stack components. Deploying and managing MLOps tools is tricky 😭😵‍💫 It is not trivial to set up all the different tools that you might need for your pipeline. 🌈 Each tool comes with a certain set of requirements. For example, a Kubeflow installation will require you to have a Kubernetes cluster, and so would a Seldon Core deployment. 🤔 Figuring out the defaults for infra parameters is not easy. Even if you have identified the backing infra that you need for a stack component, setting up reasonable defaults for parameters like instance size, CPU, memory, etc., needs a lot of experimentation to figure out. 🚧 Many times, standard tool installations don''t work out of the box. For example, to run a custom pipeline in Vertex AI, it is not enough to just run an imported pipeline. You might also need a custom service account that is configured to perform tasks like reading secrets from your secret store or talking to other GCP services that your pipeline might need. 🔐 Some tools need an additional layer of installations to enable a more secure, production-grade setup. For example, a standard MLflow tracking server deployment comes without an authentication frontend which might expose all of your tracking data to the world if deployed as-is.' - source_sentence: What is the expiration time for the GCP OAuth2 token in the ZenML configuration? sentences: - '━━━━━┛ Configuration ┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃ ┠────────────┼────────────┨ ┃ project_id │ zenml-core ┃ ┠────────────┼────────────┨ ┃ token │ [HIDDEN] ┃ ┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛ Note the temporary nature of the Service Connector. It will expire and become unusable in 1 hour: zenml service-connector list --name gcp-oauth2-token Example Command Output ┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓ ┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃ ┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨ ┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃ ┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃ ┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃ ┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃ ┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛ Auto-configuration The GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.' - 'Hugging Face Deploying models to Huggingface Inference Endpoints with Hugging Face :hugging_face:. Hugging Face Inference Endpoints provides a secure production solution to easily deploy any transformers, sentence-transformers, and diffusers models on a dedicated and autoscaling infrastructure managed by Hugging Face. An Inference Endpoint is built from a model from the Hub. This service provides dedicated and autoscaling infrastructure managed by Hugging Face, allowing you to deploy models without dealing with containers and GPUs. When to use it? You should use Hugging Face Model Deployer: if you want to deploy Transformers, Sentence-Transformers, or Diffusion models on dedicated and secure infrastructure. if you prefer a fully-managed production solution for inference without the need to handle containers and GPUs. if your goal is to turn your models into production-ready APIs with minimal infrastructure or MLOps involvement Cost-effectiveness is crucial, and you want to pay only for the raw compute resources you use. Enterprise security is a priority, and you need to deploy models into secure offline endpoints accessible only via a direct connection to your Virtual Private Cloud (VPCs). If you are looking for a more easy way to deploy your models locally, you can use the MLflow Model Deployer flavor. How to deploy it? The Hugging Face Model Deployer flavor is provided by the Hugging Face ZenML integration, so you need to install it on your local machine to be able to deploy your models. You can do this by running the following command: zenml integration install huggingface -y To register the Hugging Face model deployer with ZenML you need to run the following command: zenml model-deployer register <MODEL_DEPLOYER_NAME> --flavor=huggingface --token=<YOUR_HF_TOKEN> --namespace=<YOUR_HF_NAMESPACE> Here, token parameter is the Hugging Face authentication token. It can be managed through Hugging Face settings.' - Can you list the steps to set up a Docker registry on a Kubernetes cluster? model-index: - name: zenml/finetuned-snowflake-arctic-embed-m results: - task: type: information-retrieval name: Information Retrieval dataset: name: dim 384 type: dim_384 metrics: - type: cosine_accuracy@1 value: 0.29518072289156627 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.5240963855421686 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5843373493975904 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.6867469879518072 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.29518072289156627 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.17469879518072293 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.11686746987951804 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.0686746987951807 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.29518072289156627 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.5240963855421686 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5843373493975904 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.6867469879518072 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.4908042072911187 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.42844234079173843 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.43576329240226386 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 256 type: dim_256 metrics: - type: cosine_accuracy@1 value: 0.25903614457831325 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.5060240963855421 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5783132530120482 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.6445783132530121 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.25903614457831325 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.1686746987951807 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.11566265060240961 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.0644578313253012 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.25903614457831325 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.5060240963855421 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5783132530120482 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.6445783132530121 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.4548319777111225 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.39346194301013593 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.40343211538391555 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 128 type: dim_128 metrics: - type: cosine_accuracy@1 value: 0.2710843373493976 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.46987951807228917 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5662650602409639 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.6144578313253012 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.2710843373493976 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.1566265060240964 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.11325301204819276 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.061445783132530116 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.2710843373493976 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.46987951807228917 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5662650602409639 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.6144578313253012 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.44433019669319024 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.3893574297188756 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.3989315479842741 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 64 type: dim_64 metrics: - type: cosine_accuracy@1 value: 0.21686746987951808 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.42168674698795183 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5180722891566265 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.5843373493975904 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.21686746987951808 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.14056224899598396 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.10361445783132528 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.05843373493975902 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.21686746987951808 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.42168674698795183 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5180722891566265 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.5843373493975904 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.39639025659520544 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.3364529546758464 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.34658882510541217 name: Cosine Map@100 --- # zenml/finetuned-snowflake-arctic-embed-m This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m) <!-- at revision 71bc94c8f9ea1e54fba11167004205a65e5da2cc --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 768 tokens - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> - Language: en - License: apache-2.0 ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("zenml/finetuned-snowflake-arctic-embed-m") # Run inference sentences = [ 'What is the expiration time for the GCP OAuth2 token in the ZenML configuration?', '━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃\n\n┠────────────┼────────────┨\n\n┃ project_id │ zenml-core ┃\n\n┠────────────┼────────────┨\n\n┃ token │ [HIDDEN] ┃\n\n┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛\n\nNote the temporary nature of the Service Connector. It will expire and become unusable in 1 hour:\n\nzenml service-connector list --name gcp-oauth2-token\n\nExample Command Output\n\n┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓\n\n┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃\n\n┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨\n\n┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃\n\n┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛\n\nAuto-configuration\n\nThe GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.', 'Can you list the steps to set up a Docker registry on a Kubernetes cluster?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_384` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.2952 \| \| cosine_accuracy@3 \| 0.5241 \| \| cosine_accuracy@5 \| 0.5843 \| \| cosine_accuracy@10 \| 0.6867 \| \| cosine_precision@1 \| 0.2952 \| \| cosine_precision@3 \| 0.1747 \| \| cosine_precision@5 \| 0.1169 \| \| cosine_precision@10 \| 0.0687 \| \| cosine_recall@1 \| 0.2952 \| \| cosine_recall@3 \| 0.5241 \| \| cosine_recall@5 \| 0.5843 \| \| cosine_recall@10 \| 0.6867 \| \| cosine_ndcg@10 \| 0.4908 \| \| cosine_mrr@10 \| 0.4284 \| \| cosine_map@100 \| 0.4358 \| #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.259 \| \| cosine_accuracy@3 \| 0.506 \| \| cosine_accuracy@5 \| 0.5783 \| \| cosine_accuracy@10 \| 0.6446 \| \| cosine_precision@1 \| 0.259 \| \| cosine_precision@3 \| 0.1687 \| \| cosine_precision@5 \| 0.1157 \| \| cosine_precision@10 \| 0.0645 \| \| cosine_recall@1 \| 0.259 \| \| cosine_recall@3 \| 0.506 \| \| cosine_recall@5 \| 0.5783 \| \| cosine_recall@10 \| 0.6446 \| \| cosine_ndcg@10 \| 0.4548 \| \| cosine_mrr@10 \| 0.3935 \| \| cosine_map@100 \| 0.4034 \| #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.2711 \| \| cosine_accuracy@3 \| 0.4699 \| \| cosine_accuracy@5 \| 0.5663 \| \| cosine_accuracy@10 \| 0.6145 \| \| cosine_precision@1 \| 0.2711 \| \| cosine_precision@3 \| 0.1566 \| \| cosine_precision@5 \| 0.1133 \| \| cosine_precision@10 \| 0.0614 \| \| cosine_recall@1 \| 0.2711 \| \| cosine_recall@3 \| 0.4699 \| \| cosine_recall@5 \| 0.5663 \| \| cosine_recall@10 \| 0.6145 \| \| cosine_ndcg@10 \| 0.4443 \| \| cosine_mrr@10 \| 0.3894 \| \| cosine_map@100 \| 0.3989 \| #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.2169 \| \| cosine_accuracy@3 \| 0.4217 \| \| cosine_accuracy@5 \| 0.5181 \| \| cosine_accuracy@10 \| 0.5843 \| \| cosine_precision@1 \| 0.2169 \| \| cosine_precision@3 \| 0.1406 \| \| cosine_precision@5 \| 0.1036 \| \| cosine_precision@10 \| 0.0584 \| \| cosine_recall@1 \| 0.2169 \| \| cosine_recall@3 \| 0.4217 \| \| cosine_recall@5 \| 0.5181 \| \| cosine_recall@10 \| 0.5843 \| \| cosine_ndcg@10 \| 0.3964 \| \| cosine_mrr@10 \| 0.3365 \| \| cosine_map@100 \| 0.3466 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 1,490 training samples * Columns: <code>positive</code>, <code>anchor</code>, and <code>negative</code> * Approximate statistics based on the first 1000 samples: \| \| positive \| anchor \| negative \| \|:--------\|:----------------------------------------------------------------------------------\|:-------------------------------------------------------------------------------------\|:-----------------------------------------------------------------------------------\| \| type \| string \| string \| string \| \| details \| <ul><li>min: 9 tokens</li><li>mean: 21.02 tokens</li><li>max: 64 tokens</li></ul> \| <ul><li>min: 23 tokens</li><li>mean: 375.16 tokens</li><li>max: 512 tokens</li></ul> \| <ul><li>min: 10 tokens</li><li>mean: 17.51 tokens</li><li>max: 31 tokens</li></ul> \| * Samples: \| positive \| anchor \| negative \| \|:-----------------------------------------------------------------------------------------------------------------------------------\|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\|:-----------------------------------------------------------------------------------------------------\| \| <code>What details can you provide about the mlflow_training_pipeline runs listed in the ZenML documentation?</code> \| <code>mlflow_training_pipeline', ┃┃ │ │ │ 'zenml_pipeline_run_uuid': 'a5d4faae-ef70-48f2-9893-6e65d5e51e98', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.005'} ┃<br><br>┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨<br><br>┃ tensorflow-mnist-model │ 2 │ Run #2 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_09_08_467212', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃<br><br>┃ │ │ │ 'zenml_pipeline_run_uuid': '11858dcf-3e47-4b1a-82c5-6fa25ba4e037', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.003'} ┃<br><br>┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨<br><br>┃ tensorflow-mnist-model │ 1 │ Run #1 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_08_52_398499', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃<br><br>┃ │ │ │ 'zenml_pipeline_run_uuid': '29fb22c1-6e0b-4431-9e04-226226506d16', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.001'} ┃</code> \| <code>Can you explain how to configure the TensorFlow settings for a different project?</code> \| \| <code>How do you register a GCP Service Connector that uses account impersonation to access the zenml-bucket-sl GCS bucket?</code> \| <code>esource-id zenml-bucket-sl<br><br>Example Command OutputError: Service connector 'gcp-empty-sa' verification failed: connector authorization failure: failed to fetch GCS bucket<br><br>zenml-bucket-sl: 403 GET https://storage.googleapis.com/storage/v1/b/zenml-bucket-sl?projection=noAcl&prettyPrint=false:<br><br>[email protected] does not have storage.buckets.get access to the Google Cloud Storage bucket.<br><br>Permission 'storage.buckets.get' denied on resource (or it may not exist).<br><br>Next, we'll register a GCP Service Connector that actually uses account impersonation to access the zenml-bucket-sl GCS bucket and verify that it can actually access the bucket:<br><br>zenml service-connector register gcp-impersonate-sa --type gcp --auth-method impersonation --service_account_json=@[email protected] --project_id=zenml-core --target_principal=zenml-bucket-sl@zenml-core.iam.gserviceaccount.com --resource-type gcs-bucket --resource-id gs://zenml-bucket-sl<br><br>Example Command Output<br><br>Expanding argument value service_account_json to contents of file /home/stefan/aspyre/src/zenml/[email protected].<br><br>Successfully registered service connector `gcp-impersonate-sa` with access to the following resources:<br><br>┏━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━┓<br><br>┃ RESOURCE TYPE │ RESOURCE NAMES ┃<br><br>┠───────────────┼──────────────────────┨<br><br>┃ 📦 gcs-bucket │ gs://zenml-bucket-sl ┃<br><br>┗━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━┛<br><br>External Account (GCP Workload Identity)<br><br>Use GCP workload identity federation to authenticate to GCP services using AWS IAM credentials, Azure Active Directory credentials or generic OIDC tokens.</code> \| <code>What is the process for setting up a ZenML pipeline using AWS IAM credentials?</code> \| \| <code>Can you explain how data validation helps in detecting data drift and model drift in ZenML pipelines?</code> \| <code>of your models at different stages of development.if you have pipelines that regularly ingest new data, you should use data validation to run regular data integrity checks to signal problems before they are propagated downstream.<br><br>in continuous training pipelines, you should use data validation techniques to compare new training data against a data reference and to compare the performance of newly trained models against previous ones.<br><br>when you have pipelines that automate batch inference or if you regularly collect data used as input in online inference, you should use data validation to run data drift analyses and detect training-serving skew, data drift and model drift.<br><br>Data Validator Flavors<br><br>Data Validator are optional stack components provided by integrations. The following table lists the currently available Data Validators and summarizes their features and the data types and model types that they can be used with in ZenML pipelines:<br><br>Data Validator Validation Features Data Types Model Types Notes Flavor/Integration Deepchecks data quality<br>data drift<br>model drift<br>model performance tabular: pandas.DataFrame CV: torch.utils.data.dataloader.DataLoader tabular: sklearn.base.ClassifierMixin CV: torch.nn.Module Add Deepchecks data and model validation tests to your pipelines deepchecks Evidently data quality<br>data drift<br>model drift<br>model performance tabular: pandas.DataFrame N/A Use Evidently to generate a variety of data quality and data/model drift reports and visualizations evidently Great Expectations data profiling<br>data quality tabular: pandas.DataFrame N/A Perform data testing, documentation and profiling with Great Expectations great_expectations Whylogs/WhyLabs data drift tabular: pandas.DataFrame N/A Generate data profiles with whylogs and upload them to WhyLabs whylogs<br><br>If you would like to see the available flavors of Data Validator, you can use the command:<br><br>zenml data-validator flavor list<br><br>How to use it</code> \| <code>What are the best practices for deploying web applications using Docker and Kubernetes?</code> \| * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "TripletLoss", "matryoshka_dims": [ 384, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `gradient_accumulation_steps`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 4 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 16 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 4 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: True - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs \| Epoch \| Step \| dim_128_cosine_map@100 \| dim_256_cosine_map@100 \| dim_384_cosine_map@100 \| dim_64_cosine_map@100 \| \|:-------:\|:-----:\|:----------------------:\|:----------------------:\|:----------------------:\|:---------------------:\| \| 0.6667 \| 1 \| 0.3884 \| 0.4332 \| 0.4464 \| 0.3140 \| \| 2.0 \| 3 \| 0.4064 \| 0.4195 \| 0.4431 \| 0.3553 \| \| 2.6667 \| 4 \| 0.3989 \| 0.4034 \| 0.4358 \| 0.3466 \| * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.3.1+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### TripletLoss ```bibtex @misc{hermans2017defense, title={In Defense of the Triplet Loss for Person Re-Identification}, author={Alexander Hermans and Lucas Beyer and Bastian Leibe}, year={2017}, eprint={1703.07737}, archivePrefix={arXiv}, primaryClass={cs.CV} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# zenml/finetuned-snowflake-arctic-embed-m This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m) <!-- at revision 71bc94c8f9ea1e54fba11167004205a65e5da2cc --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 768 tokens - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> - Language: en - License: apache-2.0 ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("zenml/finetuned-snowflake-arctic-embed-m") # Run inference sentences = [ 'What is the expiration time for the GCP OAuth2 token in the ZenML configuration?', '━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃\n\n┠────────────┼────────────┨\n\n┃ project_id │ zenml-core ┃\n\n┠────────────┼────────────┨\n\n┃ token │ [HIDDEN] ┃\n\n┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛\n\nNote the temporary nature of the Service Connector. It will expire and become unusable in 1 hour:\n\nzenml service-connector list --name gcp-oauth2-token\n\nExample Command Output\n\n┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓\n\n┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃\n\n┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨\n\n┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃\n\n┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛\n\nAuto-configuration\n\nThe GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.', 'Can you list the steps to set up a Docker registry on a Kubernetes cluster?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_384` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.2952 \| \| cosine_accuracy@3 \| 0.5241 \| \| cosine_accuracy@5 \| 0.5843 \| \| cosine_accuracy@10 \| 0.6867 \| \| cosine_precision@1 \| 0.2952 \| \| cosine_precision@3 \| 0.1747 \| \| cosine_precision@5 \| 0.1169 \| \| cosine_precision@10 \| 0.0687 \| \| cosine_recall@1 \| 0.2952 \| \| cosine_recall@3 \| 0.5241 \| \| cosine_recall@5 \| 0.5843 \| \| cosine_recall@10 \| 0.6867 \| \| cosine_ndcg@10 \| 0.4908 \| \| cosine_mrr@10 \| 0.4284 \| \| cosine_map@100 \| 0.4358 \| #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.259 \| \| cosine_accuracy@3 \| 0.506 \| \| cosine_accuracy@5 \| 0.5783 \| \| cosine_accuracy@10 \| 0.6446 \| \| cosine_precision@1 \| 0.259 \| \| cosine_precision@3 \| 0.1687 \| \| cosine_precision@5 \| 0.1157 \| \| cosine_precision@10 \| 0.0645 \| \| cosine_recall@1 \| 0.259 \| \| cosine_recall@3 \| 0.506 \| \| cosine_recall@5 \| 0.5783 \| \| cosine_recall@10 \| 0.6446 \| \| cosine_ndcg@10 \| 0.4548 \| \| cosine_mrr@10 \| 0.3935 \| \| cosine_map@100 \| 0.4034 \| #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.2711 \| \| cosine_accuracy@3 \| 0.4699 \| \| cosine_accuracy@5 \| 0.5663 \| \| cosine_accuracy@10 \| 0.6145 \| \| cosine_precision@1 \| 0.2711 \| \| cosine_precision@3 \| 0.1566 \| \| cosine_precision@5 \| 0.1133 \| \| cosine_precision@10 \| 0.0614 \| \| cosine_recall@1 \| 0.2711 \| \| cosine_recall@3 \| 0.4699 \| \| cosine_recall@5 \| 0.5663 \| \| cosine_recall@10 \| 0.6145 \| \| cosine_ndcg@10 \| 0.4443 \| \| cosine_mrr@10 \| 0.3894 \| \| cosine_map@100 \| 0.3989 \| #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.2169 \| \| cosine_accuracy@3 \| 0.4217 \| \| cosine_accuracy@5 \| 0.5181 \| \| cosine_accuracy@10 \| 0.5843 \| \| cosine_precision@1 \| 0.2169 \| \| cosine_precision@3 \| 0.1406 \| \| cosine_precision@5 \| 0.1036 \| \| cosine_precision@10 \| 0.0584 \| \| cosine_recall@1 \| 0.2169 \| \| cosine_recall@3 \| 0.4217 \| \| cosine_recall@5 \| 0.5181 \| \| cosine_recall@10 \| 0.5843 \| \| cosine_ndcg@10 \| 0.3964 \| \| cosine_mrr@10 \| 0.3365 \| \| cosine_map@100 \| 0.3466 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 1,490 training samples * Columns: <code>positive</code>, <code>anchor</code>, and <code>negative</code> * Approximate statistics based on the first 1000 samples: \| \| positive \| anchor \| negative \| \|:--------\|:----------------------------------------------------------------------------------\|:-------------------------------------------------------------------------------------\|:-----------------------------------------------------------------------------------\| \| type \| string \| string \| string \| \| details \| <ul><li>min: 9 tokens</li><li>mean: 21.02 tokens</li><li>max: 64 tokens</li></ul> \| <ul><li>min: 23 tokens</li><li>mean: 375.16 tokens</li><li>max: 512 tokens</li></ul> \| <ul><li>min: 10 tokens</li><li>mean: 17.51 tokens</li><li>max: 31 tokens</li></ul> \| * Samples: \| positive \| anchor \| negative \| \|:-----------------------------------------------------------------------------------------------------------------------------------\|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\|:-----------------------------------------------------------------------------------------------------\| \| <code>What details can you provide about the mlflow_training_pipeline runs listed in the ZenML documentation?</code> \| <code>mlflow_training_pipeline', ┃┃ │ │ │ 'zenml_pipeline_run_uuid': 'a5d4faae-ef70-48f2-9893-6e65d5e51e98', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.005'} ┃<br><br>┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨<br><br>┃ tensorflow-mnist-model │ 2 │ Run #2 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_09_08_467212', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃<br><br>┃ │ │ │ 'zenml_pipeline_run_uuid': '11858dcf-3e47-4b1a-82c5-6fa25ba4e037', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.003'} ┃<br><br>┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨<br><br>┃ tensorflow-mnist-model │ 1 │ Run #1 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_08_52_398499', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃<br><br>┃ │ │ │ 'zenml_pipeline_run_uuid': '29fb22c1-6e0b-4431-9e04-226226506d16', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.001'} ┃</code> \| <code>Can you explain how to configure the TensorFlow settings for a different project?</code> \| \| <code>How do you register a GCP Service Connector that uses account impersonation to access the zenml-bucket-sl GCS bucket?</code> \| <code>esource-id zenml-bucket-sl<br><br>Example Command OutputError: Service connector 'gcp-empty-sa' verification failed: connector authorization failure: failed to fetch GCS bucket<br><br>zenml-bucket-sl: 403 GET https://storage.googleapis.com/storage/v1/b/zenml-bucket-sl?projection=noAcl&prettyPrint=false:<br><br>[email protected] does not have storage.buckets.get access to the Google Cloud Storage bucket.<br><br>Permission 'storage.buckets.get' denied on resource (or it may not exist).<br><br>Next, we'll register a GCP Service Connector that actually uses account impersonation to access the zenml-bucket-sl GCS bucket and verify that it can actually access the bucket:<br><br>zenml service-connector register gcp-impersonate-sa --type gcp --auth-method impersonation --service_account_json=@[email protected] --project_id=zenml-core --target_principal=zenml-bucket-sl@zenml-core.iam.gserviceaccount.com --resource-type gcs-bucket --resource-id gs://zenml-bucket-sl<br><br>Example Command Output<br><br>Expanding argument value service_account_json to contents of file /home/stefan/aspyre/src/zenml/[email protected].<br><br>Successfully registered service connector `gcp-impersonate-sa` with access to the following resources:<br><br>┏━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━┓<br><br>┃ RESOURCE TYPE │ RESOURCE NAMES ┃<br><br>┠───────────────┼──────────────────────┨<br><br>┃ 📦 gcs-bucket │ gs://zenml-bucket-sl ┃<br><br>┗━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━┛<br><br>External Account (GCP Workload Identity)<br><br>Use GCP workload identity federation to authenticate to GCP services using AWS IAM credentials, Azure Active Directory credentials or generic OIDC tokens.</code> \| <code>What is the process for setting up a ZenML pipeline using AWS IAM credentials?</code> \| \| <code>Can you explain how data validation helps in detecting data drift and model drift in ZenML pipelines?</code> \| <code>of your models at different stages of development.if you have pipelines that regularly ingest new data, you should use data validation to run regular data integrity checks to signal problems before they are propagated downstream.<br><br>in continuous training pipelines, you should use data validation techniques to compare new training data against a data reference and to compare the performance of newly trained models against previous ones.<br><br>when you have pipelines that automate batch inference or if you regularly collect data used as input in online inference, you should use data validation to run data drift analyses and detect training-serving skew, data drift and model drift.<br><br>Data Validator Flavors<br><br>Data Validator are optional stack components provided by integrations. The following table lists the currently available Data Validators and summarizes their features and the data types and model types that they can be used with in ZenML pipelines:<br><br>Data Validator Validation Features Data Types Model Types Notes Flavor/Integration Deepchecks data quality<br>data drift<br>model drift<br>model performance tabular: pandas.DataFrame CV: torch.utils.data.dataloader.DataLoader tabular: sklearn.base.ClassifierMixin CV: torch.nn.Module Add Deepchecks data and model validation tests to your pipelines deepchecks Evidently data quality<br>data drift<br>model drift<br>model performance tabular: pandas.DataFrame N/A Use Evidently to generate a variety of data quality and data/model drift reports and visualizations evidently Great Expectations data profiling<br>data quality tabular: pandas.DataFrame N/A Perform data testing, documentation and profiling with Great Expectations great_expectations Whylogs/WhyLabs data drift tabular: pandas.DataFrame N/A Generate data profiles with whylogs and upload them to WhyLabs whylogs<br><br>If you would like to see the available flavors of Data Validator, you can use the command:<br><br>zenml data-validator flavor list<br><br>How to use it</code> \| <code>What are the best practices for deploying web applications using Docker and Kubernetes?</code> \| * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "TripletLoss", "matryoshka_dims": [ 384, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `gradient_accumulation_steps`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 4 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 16 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 4 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: True - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs \| Epoch \| Step \| dim_128_cosine_map@100 \| dim_256_cosine_map@100 \| dim_384_cosine_map@100 \| dim_64_cosine_map@100 \| \|:-------:\|:-----:\|:----------------------:\|:----------------------:\|:----------------------:\|:---------------------:\| \| 0.6667 \| 1 \| 0.3884 \| 0.4332 \| 0.4464 \| 0.3140 \| \| 2.0 \| 3 \| 0.4064 \| 0.4195 \| 0.4431 \| 0.3553 \| \| 2.6667 \| 4 \| 0.3989 \| 0.4034 \| 0.4358 \| 0.3466 \| * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.3.1+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### TripletLoss ```bibtex @misc{hermans2017defense, title={In Defense of the Triplet Loss for Person Re-Identification}, author={Alexander Hermans and Lucas Beyer and Bastian Leibe}, year={2017}, eprint={1703.07737}, archivePrefix={arXiv}, primaryClass={cs.CV} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "Snowflake/snowflake-arctic-embed-m", "datasets": [], "language": ["en"], "library_name": "sentence-transformers", "license": "apache-2.0", "metrics": ["cosine_accuracy@1", "cosine_accuracy@3", "cosine_accuracy@5", "cosine_accuracy@10", "cosine_precision@1", "cosine_precision@3", "cosine_precision@5", "cosine_precision@10", "cosine_recall@1", "cosine_recall@3", "cosine_recall@5", "cosine_recall@10", "cosine_ndcg@10", "cosine_mrr@10", "cosine_map@100"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:1490", "loss:MatryoshkaLoss", "loss:TripletLoss"], "widget": [{"source_sentence": "Where is the global configuration directory located in ZenML's default setup?", "sentences": ["'default' ...\n\nCreating default user 'default' ...Creating default stack for user 'default' in workspace default...\n\nActive workspace not set. Setting it to the default.\n\nThe active stack is not set. Setting the active stack to the default workspace stack.\n\nUsing the default store for the global config.\n\nUnable to find ZenML repository in your current working directory (/tmp/folder) or any parent directories. If you want to use an existing repository which is in a different location, set the environment variable 'ZENML_REPOSITORY_PATH'. If you want to create a new repository, run zenml init.\n\nRunning without an active repository root.\n\nUsing the default local database.\n\nRunning with active workspace: 'default' (global)\n\n┏━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━┓\n\n┃ ACTIVE │ STACK NAME │ SHARED │ OWNER │ ARTIFACT_STORE │ ORCHESTRATOR ┃\n\n┠────────┼────────────┼────────┼─────────┼────────────────┼──────────────┨\n\n┃ 👉 │ default │ ❌ │ default │ default │ default ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━┛\n\nThe following is an example of the layout of the global config directory immediately after initialization:\n\n/home/stefan/.config/zenml <- Global Config Directory\n\n├── config.yaml <- Global Configuration Settings\n\n└── local_stores <- Every Stack component that stores information\n\n\| locally will have its own subdirectory here.\n\n├── a1a0d3d0-d552-4a80-be09-67e5e29be8ee <- e.g. Local Store path for the\n\n\| `default` local Artifact Store\n\n└── default_zen_store\n\n└── zenml.db <- SQLite database where ZenML data (stacks,\n\ncomponents, etc) are stored by default.\n\nAs shown above, the global config directory stores the following information:", "How do you configure the network settings on a Linux server?", "Reranking for better retrieval\n\nAdd reranking to your RAG inference for better retrieval performance.\n\nRerankers are a crucial component of retrieval systems that use LLMs. They help improve the quality of the retrieved documents by reordering them based on additional features or scores. In this section, we'll explore how to add a reranker to your RAG inference pipeline in ZenML.\n\nIn previous sections, we set up the overall workflow, from data ingestion and preprocessing to embeddings generation and retrieval. We then set up some basic evaluation metrics to assess the performance of our retrieval system. A reranker is a way to squeeze a bit of extra performance out of the system by reordering the retrieved documents based on additional features or scores.\n\nAs you can see, reranking is an optional addition we make to what we've already set up. It's not strictly necessary, but it can help improve the relevance and quality of the retrieved documents, which in turn can lead to better responses from the LLM. Let's dive in!\n\nPreviousEvaluation in practice\n\nNextUnderstanding reranking\n\nLast updated 1 month ago"]}, {"source_sentence": "Where can I find the instructions to enable CUDA for GPU-backed hardware in ZenML SDK Docs?", "sentences": ["Migration guide 0.39.1 → 0.41.0\n\nHow to migrate your ZenML pipelines and steps from version <=0.39.1 to 0.41.0.\n\nZenML versions 0.40.0 to 0.41.0 introduced a new and more flexible syntax to define ZenML steps and pipelines. This page contains code samples that show you how to upgrade your steps and pipelines to the new syntax.\n\nNewer versions of ZenML still work with pipelines and steps defined using the old syntax, but the old syntax is deprecated and will be removed in the future.\n\nOverview\n\nfrom typing import Optional\n\nfrom zenml.steps import BaseParameters, Output, StepContext, step\n\nfrom zenml.pipelines import pipeline\n\n# Define a Step\n\nclass MyStepParameters(BaseParameters):\n\nparam_1: int\n\nparam_2: Optional[float] = None\n\n@step\n\ndef my_step(\n\nparams: MyStepParameters, context: StepContext,\n\n) -> Output(int_output=int, str_output=str):\n\nresult = int(params.param_1 * (params.param_2 or 1))\n\nresult_uri = context.get_output_artifact_uri()\n\nreturn result, result_uri\n\n# Run the Step separately\n\nmy_step.entrypoint()\n\n# Define a Pipeline\n\n@pipeline\n\ndef my_pipeline(my_step):\n\nmy_step()\n\nstep_instance = my_step(params=MyStepParameters(param_1=17))\n\npipeline_instance = my_pipeline(my_step=step_instance)\n\n# Configure and run the Pipeline\n\npipeline_instance.configure(enable_cache=False)\n\nschedule = Schedule(...)\n\npipeline_instance.run(schedule=schedule)\n\n# Fetch the Pipeline Run\n\nlast_run = pipeline_instance.get_runs()[0]\n\nint_output = last_run.get_step[\"my_step\"].outputs[\"int_output\"].read()\n\nfrom typing import Annotated, Optional, Tuple\n\nfrom zenml import get_step_context, pipeline, step\n\nfrom zenml.client import Client\n\n# Define a Step\n\n@step\n\ndef my_step(\n\nparam_1: int, param_2: Optional[float] = None\n\n) -> Tuple[Annotated[int, \"int_output\"], Annotated[str, \"str_output\"]]:\n\nresult = int(param_1 * (param_2 or 1))\n\nresult_uri = get_step_context().get_output_artifact_uri()\n\nreturn result, result_uri\n\n# Run the Step separately\n\nmy_step()\n\n# Define a Pipeline\n\n@pipeline", "How do I integrate Google Cloud VertexAI into my existing Kubernetes cluster?", " SDK Docs .\n\nEnabling CUDA for GPU-backed hardwareNote that if you wish to use this step operator to run steps on a GPU, you will need to follow the instructions on this page to ensure that it works. It requires adding some extra settings customization and is essential to enable CUDA for the GPU to give its full acceleration.\n\nPreviousStep Operators\n\nNextGoogle Cloud VertexAI\n\nLast updated 19 days ago"]}, {"source_sentence": "What are the special metadata types supported by ZenML and how are they used?", "sentences": ["Special Metadata Types\n\nTracking your metadata.\n\nZenML supports several special metadata types to capture specific kinds of information. Here are examples of how to use the special types Uri, Path, DType, and StorageSize:\n\nfrom zenml.metadata.metadata_types import StorageSize, DType\n\nfrom zenml import log_artifact_metadata\n\nlog_artifact_metadata(\n\nmetadata={\n\n\"dataset_source\": Uri(\"gs://my-bucket/datasets/source.csv\"),\n\n\"preprocessing_script\": Path(\"/scripts/preprocess.py\"),\n\n\"column_types\": {\n\n\"age\": DType(\"int\"),\n\n\"income\": DType(\"float\"),\n\n\"score\": DType(\"int\")\n\n},\n\n\"processed_data_size\": StorageSize(2500000)\n\nIn this example:\n\nUri is used to indicate a dataset source URI.\n\nPath is used to specify the filesystem path to a preprocessing script.\n\nDType is used to describe the data types of specific columns.\n\nStorageSize is used to indicate the size of the processed data in bytes.\n\nThese special types help standardize the format of metadata and ensure that it is logged in a consistent and interpretable manner.\n\nPreviousGroup metadata\n\nNextFetch metadata within steps\n\nLast updated 19 days ago", "Configure a code repository\n\nConnect a Git repository to ZenML to track code changes and collaborate on MLOps projects.\n\nThroughout the lifecycle of a MLOps pipeline, it can get quite tiresome to always wait for a Docker build every time after running a pipeline (even if the local Docker cache is used). However, there is a way to just have one pipeline build and keep reusing it until a change to the pipeline environment is made: by connecting a code repository.\n\nWith ZenML, connecting to a Git repository optimizes the Docker build processes. It also has the added bonus of being a better way of managing repository changes and enabling better code collaboration. Here is how the flow changes when running a pipeline:\n\nYou trigger a pipeline run on your local machine. ZenML parses the @pipeline function to determine the necessary steps.\n\nThe local client requests stack information from the ZenML server, which responds with the cloud stack configuration.\n\nThe local client detects that we're using a code repository and requests the information from the git repo.\n\nInstead of building a new Docker image, the client checks if an existing image can be reused based on the current Git commit hash and other environment metadata.\n\nThe client initiates a run in the orchestrator, which sets up the execution environment in the cloud, such as a VM.\n\nThe orchestrator downloads the code directly from the Git repository and uses the existing Docker image to run the pipeline steps.\n\nPipeline steps execute, storing artifacts in the cloud-based artifact store.\n\nThroughout the execution, the pipeline run status and metadata are reported back to the ZenML server.\n\nBy connecting a Git repository, you avoid redundant builds and make your MLOps processes more efficient. Your team can work on the codebase simultaneously, with ZenML handling the version tracking and ensuring that the correct code version is always used for each run.\n\nCreating a GitHub Repository", "Can you explain the process of setting up a virtual environment in Python?"]}, {"source_sentence": "What are the benefits of deploying stack components directly from the ZenML CLI?", "sentences": ["─────────────────────────────────────────────────┨┃ RESOURCE TYPES │ 🔵 gcp-generic, 📦 gcs-bucket, 🌀 kubernetes-cluster, 🐳 docker-registry ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ RESOURCE NAME │ <multiple> ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ SECRET ID │ 4694de65-997b-4929-8831-b49d5e067b97 ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ SESSION DURATION │ N/A ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ EXPIRES IN │ 59m46s ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ OWNER │ default ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ WORKSPACE │ default ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ SHARED │ ➖ ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ CREATED_AT │ 2023-05-19 09:04:33.557126 ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ UPDATED_AT │ 2023-05-19 09:04:33.557127 ┃\n\n┗━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓", "How do you set up a custom service account for Vertex AI?", "⚒️Manage stacks\n\nDeploying your stack components directly from the ZenML CLI\n\nThe first step in running your pipelines on remote infrastructure is to deploy all the components that you would need, like an MLflow tracking server, a Seldon Core model deployer, and more to your cloud.\n\nThis can bring plenty of benefits like scalability, reliability, and collaboration. ZenML eases the path to production by providing a seamless way for all tools to interact with others through the use of abstractions. However, one of the most painful parts of this process, from what we see on our Slack and in general, is the deployment of these stack components.\n\nDeploying and managing MLOps tools is tricky 😭😵‍💫\n\nIt is not trivial to set up all the different tools that you might need for your pipeline.\n\n🌈 Each tool comes with a certain set of requirements. For example, a Kubeflow installation will require you to have a Kubernetes cluster, and so would a Seldon Core deployment.\n\n🤔 Figuring out the defaults for infra parameters is not easy. Even if you have identified the backing infra that you need for a stack component, setting up reasonable defaults for parameters like instance size, CPU, memory, etc., needs a lot of experimentation to figure out.\n\n🚧 Many times, standard tool installations don't work out of the box. For example, to run a custom pipeline in Vertex AI, it is not enough to just run an imported pipeline. You might also need a custom service account that is configured to perform tasks like reading secrets from your secret store or talking to other GCP services that your pipeline might need.\n\n🔐 Some tools need an additional layer of installations to enable a more secure, production-grade setup. For example, a standard MLflow tracking server deployment comes without an authentication frontend which might expose all of your tracking data to the world if deployed as-is."]}, {"source_sentence": "What is the expiration time for the GCP OAuth2 token in the ZenML configuration?", "sentences": ["━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃\n\n┠────────────┼────────────┨\n\n┃ project_id │ zenml-core ┃\n\n┠────────────┼────────────┨\n\n┃ token │ [HIDDEN] ┃\n\n┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛\n\nNote the temporary nature of the Service Connector. It will expire and become unusable in 1 hour:\n\nzenml service-connector list --name gcp-oauth2-token\n\nExample Command Output\n\n┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓\n\n┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃\n\n┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨\n\n┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃\n\n┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛\n\nAuto-configuration\n\nThe GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.", "Hugging Face\n\nDeploying models to Huggingface Inference Endpoints with Hugging Face :hugging_face:.\n\nHugging Face Inference Endpoints provides a secure production solution to easily deploy any transformers, sentence-transformers, and diffusers models on a dedicated and autoscaling infrastructure managed by Hugging Face. An Inference Endpoint is built from a model from the Hub.\n\nThis service provides dedicated and autoscaling infrastructure managed by Hugging Face, allowing you to deploy models without dealing with containers and GPUs.\n\nWhen to use it?\n\nYou should use Hugging Face Model Deployer:\n\nif you want to deploy Transformers, Sentence-Transformers, or Diffusion models on dedicated and secure infrastructure.\n\nif you prefer a fully-managed production solution for inference without the need to handle containers and GPUs.\n\nif your goal is to turn your models into production-ready APIs with minimal infrastructure or MLOps involvement\n\nCost-effectiveness is crucial, and you want to pay only for the raw compute resources you use.\n\nEnterprise security is a priority, and you need to deploy models into secure offline endpoints accessible only via a direct connection to your Virtual Private Cloud (VPCs).\n\nIf you are looking for a more easy way to deploy your models locally, you can use the MLflow Model Deployer flavor.\n\nHow to deploy it?\n\nThe Hugging Face Model Deployer flavor is provided by the Hugging Face ZenML integration, so you need to install it on your local machine to be able to deploy your models. You can do this by running the following command:\n\nzenml integration install huggingface -y\n\nTo register the Hugging Face model deployer with ZenML you need to run the following command:\n\nzenml model-deployer register <MODEL_DEPLOYER_NAME> --flavor=huggingface --token=<YOUR_HF_TOKEN> --namespace=<YOUR_HF_NAMESPACE>\n\nHere,\n\ntoken parameter is the Hugging Face authentication token. It can be managed through Hugging Face settings.", "Can you list the steps to set up a Docker registry on a Kubernetes cluster?"]}], "model-index": [{"name": "zenml/finetuned-snowflake-arctic-embed-m", "results": [{"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 384", "type": "dim_384"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.29518072289156627, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.5240963855421686, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5843373493975904, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.6867469879518072, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.29518072289156627, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.17469879518072293, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.11686746987951804, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.0686746987951807, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.29518072289156627, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.5240963855421686, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5843373493975904, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.6867469879518072, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.4908042072911187, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.42844234079173843, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.43576329240226386, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 256", "type": "dim_256"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.25903614457831325, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.5060240963855421, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5783132530120482, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.6445783132530121, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.25903614457831325, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.1686746987951807, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.11566265060240961, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.0644578313253012, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.25903614457831325, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.5060240963855421, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5783132530120482, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.6445783132530121, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.4548319777111225, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.39346194301013593, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.40343211538391555, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 128", "type": "dim_128"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.2710843373493976, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.46987951807228917, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5662650602409639, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.6144578313253012, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.2710843373493976, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.1566265060240964, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.11325301204819276, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.061445783132530116, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.2710843373493976, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.46987951807228917, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5662650602409639, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.6144578313253012, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.44433019669319024, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.3893574297188756, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.3989315479842741, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 64", "type": "dim_64"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.21686746987951808, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.42168674698795183, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5180722891566265, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.5843373493975904, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.21686746987951808, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.14056224899598396, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.10361445783132528, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.05843373493975902, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.21686746987951808, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.42168674698795183, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5180722891566265, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.5843373493975904, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.39639025659520544, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.3364529546758464, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.34658882510541217, "name": "Cosine Map@100"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,574
debarghabhattofficial/t5-small-squad-qg-a2c-spt-valid	debarghabhattofficial	text2text-generation	[ "transformers", "pytorch", "tensorboard", "t5", "text2text-generation", "generated_from_trainer", "dataset:qg_squad", "license:cc-by-4.0", "model-index", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2023-04-03T11:39:31Z	2023-04-03T19:25:27+00:00	12	0	--- datasets: - qg_squad license: cc-by-4.0 metrics: - bleu tags: - generated_from_trainer model-index: - name: t5-small-squad-qg-a2c-spt-valid results: - task: type: text2text-generation name: Sequence-to-sequence Language Modeling dataset: name: qg_squad type: qg_squad config: qg_squad split: test args: qg_squad metrics: - type: bleu value: 0.1856298695745541 name: Bleu --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # t5-small-squad-qg-a2c-spt-valid This model is a fine-tuned version of [lmqg/t5-small-squad-qg](https://huggingface.co/lmqg/t5-small-squad-qg) on the qg_squad dataset. It achieves the following results on the evaluation set: - Loss: 3.5585 - Bleu: 0.1856 - Precisions: [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] - Brevity Penalty: 0.9231 - Length Ratio: 0.9259 - Translation Length: 126899 - Reference Length: 137056 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 - label_smoothing_factor: 0.15 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Bleu \| Precisions \| Brevity Penalty \| Length Ratio \| Translation Length \| Reference Length \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:------:\|:-----------------------------------------------------------------------------------:\|:---------------:\|:------------:\|:------------------:\|:----------------:\| \| 3.4717 \| 1.0 \| 1184 \| 3.5703 \| 0.1850 \| [0.4884210026960997, 0.23740423378300554, 0.14702360671696277, 0.09591845720324058] \| 0.9198 \| 0.9228 \| 126479 \| 137056 \| \| 3.4432 \| 2.0 \| 2368 \| 3.5676 \| 0.1847 \| [0.4899809765377299, 0.23739313808702955, 0.14709099076226004, 0.09610180163262601] \| 0.9173 \| 0.9205 \| 126160 \| 137056 \| \| 3.4207 \| 3.0 \| 3552 \| 3.5654 \| 0.1855 \| [0.48690609948692964, 0.236654650074526, 0.14669770766719153, 0.09533838196460138] \| 0.9260 \| 0.9286 \| 127273 \| 137056 \| \| 3.4017 \| 4.0 \| 4736 \| 3.5575 \| 0.1861 \| [0.4907433036243861, 0.23905491743183327, 0.14802083840498564, 0.09654473782730295] \| 0.9195 \| 0.9226 \| 126449 \| 137056 \| \| 3.3862 \| 5.0 \| 5920 \| 3.5540 \| 0.1851 \| [0.4916027385306181, 0.23877172085201795, 0.14769450336757936, 0.09608281170511601] \| 0.9164 \| 0.9197 \| 126053 \| 137056 \| \| 3.3715 \| 6.0 \| 7104 \| 3.5619 \| 0.1847 \| [0.4897172642552519, 0.23742624822429256, 0.14650127350144848, 0.09495653320731078] \| 0.9209 \| 0.9239 \| 126620 \| 137056 \| \| 3.3602 \| 7.0 \| 8288 \| 3.5581 \| 0.1857 \| [0.49199648336329865, 0.2390627732121, 0.14782006380301063, 0.09637410897534923] \| 0.9180 \| 0.9212 \| 126257 \| 137056 \| \| 3.3523 \| 8.0 \| 9472 \| 3.5575 \| 0.1856 \| [0.4896288812767368, 0.23802266135985578, 0.14728396021137705, 0.09588544697859817] \| 0.9215 \| 0.9244 \| 126698 \| 137056 \| \| 3.3439 \| 9.0 \| 10656 \| 3.5582 \| 0.1862 \| [0.4919672196048933, 0.23971752696254087, 0.14848694668474074, 0.09658739962940087] \| 0.9183 \| 0.9215 \| 126295 \| 137056 \| \| 3.3395 \| 10.0 \| 11840 \| 3.5585 \| 0.1856 \| [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] \| 0.9231 \| 0.9259 \| 126899 \| 137056 \| ### Framework versions - Transformers 4.27.4 - Pytorch 1.9.0 - Datasets 2.9.0 - Tokenizers 0.13.2	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # t5-small-squad-qg-a2c-spt-valid This model is a fine-tuned version of [lmqg/t5-small-squad-qg](https://huggingface.co/lmqg/t5-small-squad-qg) on the qg_squad dataset. It achieves the following results on the evaluation set: - Loss: 3.5585 - Bleu: 0.1856 - Precisions: [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] - Brevity Penalty: 0.9231 - Length Ratio: 0.9259 - Translation Length: 126899 - Reference Length: 137056 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 - label_smoothing_factor: 0.15 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Bleu \| Precisions \| Brevity Penalty \| Length Ratio \| Translation Length \| Reference Length \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:------:\|:-----------------------------------------------------------------------------------:\|:---------------:\|:------------:\|:------------------:\|:----------------:\| \| 3.4717 \| 1.0 \| 1184 \| 3.5703 \| 0.1850 \| [0.4884210026960997, 0.23740423378300554, 0.14702360671696277, 0.09591845720324058] \| 0.9198 \| 0.9228 \| 126479 \| 137056 \| \| 3.4432 \| 2.0 \| 2368 \| 3.5676 \| 0.1847 \| [0.4899809765377299, 0.23739313808702955, 0.14709099076226004, 0.09610180163262601] \| 0.9173 \| 0.9205 \| 126160 \| 137056 \| \| 3.4207 \| 3.0 \| 3552 \| 3.5654 \| 0.1855 \| [0.48690609948692964, 0.236654650074526, 0.14669770766719153, 0.09533838196460138] \| 0.9260 \| 0.9286 \| 127273 \| 137056 \| \| 3.4017 \| 4.0 \| 4736 \| 3.5575 \| 0.1861 \| [0.4907433036243861, 0.23905491743183327, 0.14802083840498564, 0.09654473782730295] \| 0.9195 \| 0.9226 \| 126449 \| 137056 \| \| 3.3862 \| 5.0 \| 5920 \| 3.5540 \| 0.1851 \| [0.4916027385306181, 0.23877172085201795, 0.14769450336757936, 0.09608281170511601] \| 0.9164 \| 0.9197 \| 126053 \| 137056 \| \| 3.3715 \| 6.0 \| 7104 \| 3.5619 \| 0.1847 \| [0.4897172642552519, 0.23742624822429256, 0.14650127350144848, 0.09495653320731078] \| 0.9209 \| 0.9239 \| 126620 \| 137056 \| \| 3.3602 \| 7.0 \| 8288 \| 3.5581 \| 0.1857 \| [0.49199648336329865, 0.2390627732121, 0.14782006380301063, 0.09637410897534923] \| 0.9180 \| 0.9212 \| 126257 \| 137056 \| \| 3.3523 \| 8.0 \| 9472 \| 3.5575 \| 0.1856 \| [0.4896288812767368, 0.23802266135985578, 0.14728396021137705, 0.09588544697859817] \| 0.9215 \| 0.9244 \| 126698 \| 137056 \| \| 3.3439 \| 9.0 \| 10656 \| 3.5582 \| 0.1862 \| [0.4919672196048933, 0.23971752696254087, 0.14848694668474074, 0.09658739962940087] \| 0.9183 \| 0.9215 \| 126295 \| 137056 \| \| 3.3395 \| 10.0 \| 11840 \| 3.5585 \| 0.1856 \| [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] \| 0.9231 \| 0.9259 \| 126899 \| 137056 \| ### Framework versions - Transformers 4.27.4 - Pytorch 1.9.0 - Datasets 2.9.0 - Tokenizers 0.13.2	{"datasets": ["qg_squad"], "license": "cc-by-4.0", "metrics": ["bleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "t5-small-squad-qg-a2c-spt-valid", "results": [{"task": {"type": "text2text-generation", "name": "Sequence-to-sequence Language Modeling"}, "dataset": {"name": "qg_squad", "type": "qg_squad", "config": "qg_squad", "split": "test", "args": "qg_squad"}, "metrics": [{"type": "bleu", "value": 0.1856298695745541, "name": "Bleu"}]}]}]}	task	[ "TRANSLATION" ]	44,575
Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF	Triangle104	text-generation	[ "transformers", "gguf", "safetensors", "onnx", "transformers.js", "llama-cpp", "gguf-my-repo", "text-generation", "en", "base_model:HuggingFaceTB/SmolLM2-135M-Instruct", "base_model:quantized:HuggingFaceTB/SmolLM2-135M-Instruct", "license:apache-2.0", "endpoints_compatible", "region:us", "conversational" ]	2024-11-07T05:02:37Z	2025-01-14T07:12:49+00:00	7	0	--- base_model: HuggingFaceTB/SmolLM2-135M-Instruct language: - en library_name: transformers license: apache-2.0 pipeline_tag: text-generation tags: - safetensors - onnx - transformers.js - llama-cpp - gguf-my-repo --- # Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF This model was converted to GGUF format from [`HuggingFaceTB/SmolLM2-135M-Instruct`](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) for more details on the model. --- Model details: - SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters. They are capable of solving a wide range of tasks while being lightweight enough to run on-device. SmolLM2 demonstrates significant advances over its predecessor SmolLM1, particularly in instruction following, knowledge, reasoning. The 135M model was trained on 2 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new filtered datasets we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback. The instruct model additionally supports tasks such as text rewriting, summarization and function calling (for the 1.7B) thanks to datasets developed by Argilla such as Synth-APIGen-v0.1. You can find the SFT dataset here: https://huggingface.co/datasets/HuggingFaceTB/smol-smoltalk and finetuning code at https://github.com/huggingface/alignment-handbook/tree/main/recipes/smollm2 How to use Transformers pip install transformers from transformers import AutoModelForCausalLM, AutoTokenizer checkpoint = "HuggingFaceTB/SmolLM2-135M-Instruct" device = "cuda" # for GPU usage or "cpu" for CPU usage tokenizer = AutoTokenizer.from_pretrained(checkpoint) # for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")` model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device) messages = [{"role": "user", "content": "What is gravity?"}] input_text=tokenizer.apply_chat_template(messages, tokenize=False) print(input_text) inputs = tokenizer.encode(input_text, return_tensors="pt").to(device) outputs = model.generate(inputs, max_new_tokens=50, temperature=0.2, top_p=0.9, do_sample=True) print(tokenizer.decode(outputs[0])) Chat in TRL You can also use the TRL CLI to chat with the model from the terminal: pip install trl trl chat --model_name_or_path HuggingFaceTB/SmolLM2-135M-Instruct --device cpu --- ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ```	null	Non_BioNLP	# Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF This model was converted to GGUF format from [`HuggingFaceTB/SmolLM2-135M-Instruct`](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) for more details on the model. --- Model details: - SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters. They are capable of solving a wide range of tasks while being lightweight enough to run on-device. SmolLM2 demonstrates significant advances over its predecessor SmolLM1, particularly in instruction following, knowledge, reasoning. The 135M model was trained on 2 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new filtered datasets we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback. The instruct model additionally supports tasks such as text rewriting, summarization and function calling (for the 1.7B) thanks to datasets developed by Argilla such as Synth-APIGen-v0.1. You can find the SFT dataset here: https://huggingface.co/datasets/HuggingFaceTB/smol-smoltalk and finetuning code at https://github.com/huggingface/alignment-handbook/tree/main/recipes/smollm2 How to use Transformers pip install transformers from transformers import AutoModelForCausalLM, AutoTokenizer checkpoint = "HuggingFaceTB/SmolLM2-135M-Instruct" device = "cuda" # for GPU usage or "cpu" for CPU usage tokenizer = AutoTokenizer.from_pretrained(checkpoint) # for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")` model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device) messages = [{"role": "user", "content": "What is gravity?"}] input_text=tokenizer.apply_chat_template(messages, tokenize=False) print(input_text) inputs = tokenizer.encode(input_text, return_tensors="pt").to(device) outputs = model.generate(inputs, max_new_tokens=50, temperature=0.2, top_p=0.9, do_sample=True) print(tokenizer.decode(outputs[0])) Chat in TRL You can also use the TRL CLI to chat with the model from the terminal: pip install trl trl chat --model_name_or_path HuggingFaceTB/SmolLM2-135M-Instruct --device cpu --- ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ```	{"base_model": "HuggingFaceTB/SmolLM2-135M-Instruct", "language": ["en"], "library_name": "transformers", "license": "apache-2.0", "pipeline_tag": "text-generation", "tags": ["safetensors", "onnx", "transformers.js", "llama-cpp", "gguf-my-repo"]}	task	[ "SUMMARIZATION" ]	44,576
fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391	fine-tuned	feature-extraction	[ "sentence-transformers", "safetensors", "bert", "feature-extraction", "sentence-similarity", "mteb", "custom_code", "en", "dataset:fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391", "dataset:allenai/c4", "license:apache-2.0", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	2024-05-28T23:43:36Z	2024-05-28T23:43:55+00:00	7	0	--- datasets: - fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391 - allenai/c4 language: - en - en license: apache-2.0 pipeline_tag: feature-extraction tags: - sentence-transformers - feature-extraction - sentence-similarity - mteb --- This model is a fine-tuned version of [jinaai/jina-embeddings-v2-base-en](https://huggingface.co/jinaai/jina-embeddings-v2-base-en) designed for the following use case: None ## How to Use This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started: ```python from sentence_transformers import SentenceTransformer from sentence_transformers.util import cos_sim model = SentenceTransformer( 'fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391', trust_remote_code=True ) embeddings = model.encode([ 'first text to embed', 'second text to embed' ]) print(cos_sim(embeddings[0], embeddings[1])) ```	null	Non_BioNLP	This model is a fine-tuned version of [jinaai/jina-embeddings-v2-base-en](https://huggingface.co/jinaai/jina-embeddings-v2-base-en) designed for the following use case: None ## How to Use This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started: ```python from sentence_transformers import SentenceTransformer from sentence_transformers.util import cos_sim model = SentenceTransformer( 'fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391', trust_remote_code=True ) embeddings = model.encode([ 'first text to embed', 'second text to embed' ]) print(cos_sim(embeddings[0], embeddings[1])) ```	{"datasets": ["fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391", "allenai/c4"], "language": ["en", "en"], "license": "apache-2.0", "pipeline_tag": "feature-extraction", "tags": ["sentence-transformers", "feature-extraction", "sentence-similarity", "mteb"]}	task	[ "TEXT_CLASSIFICATION" ]	44,577
zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF	zzz-z-z-z	translation	[ "gguf", "llama-cpp", "gguf-my-repo", "translation", "en", "de", "fr", "zh", "pt", "nl", "ru", "ko", "it", "es", "base_model:Unbabel/TowerInstruct-Mistral-7B-v0.2", "base_model:quantized:Unbabel/TowerInstruct-Mistral-7B-v0.2", "license:cc-by-nc-4.0", "endpoints_compatible", "region:us", "conversational" ]	2025-01-09T19:13:14Z	2025-01-09T19:13:50+00:00	8	0	--- base_model: Unbabel/TowerInstruct-Mistral-7B-v0.2 language: - en - de - fr - zh - pt - nl - ru - ko - it - es license: cc-by-nc-4.0 metrics: - comet pipeline_tag: translation tags: - llama-cpp - gguf-my-repo --- # zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF This model was converted to GGUF format from [`Unbabel/TowerInstruct-Mistral-7B-v0.2`](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) for more details on the model. ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ```	null	Non_BioNLP	# zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF This model was converted to GGUF format from [`Unbabel/TowerInstruct-Mistral-7B-v0.2`](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) for more details on the model. ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ```	{"base_model": "Unbabel/TowerInstruct-Mistral-7B-v0.2", "language": ["en", "de", "fr", "zh", "pt", "nl", "ru", "ko", "it", "es"], "license": "cc-by-nc-4.0", "metrics": ["comet"], "pipeline_tag": "translation", "tags": ["llama-cpp", "gguf-my-repo"]}	task	[ "TRANSLATION" ]	44,578
Xenova/opus-mt-ru-uk	Xenova	translation	[ "transformers.js", "onnx", "marian", "text2text-generation", "translation", "base_model:Helsinki-NLP/opus-mt-ru-uk", "base_model:quantized:Helsinki-NLP/opus-mt-ru-uk", "region:us" ]	2023-09-05T23:01:23Z	2024-10-08T13:41:58+00:00	57	0	--- base_model: Helsinki-NLP/opus-mt-ru-uk library_name: transformers.js pipeline_tag: translation --- https://huggingface.co/Helsinki-NLP/opus-mt-ru-uk with ONNX weights to be compatible with Transformers.js. Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`).	null	Non_BioNLP	ERROR: type should be string, got "\nhttps://huggingface.co/Helsinki-NLP/opus-mt-ru-uk with ONNX weights to be compatible with Transformers.js.\n\nNote: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`)."	{"base_model": "Helsinki-NLP/opus-mt-ru-uk", "library_name": "transformers.js", "pipeline_tag": "translation"}	task	[ "TRANSLATION" ]	44,579
EleutherAI/pile-t5-xxl	EleutherAI	text2text-generation	[ "transformers", "safetensors", "umt5", "text2text-generation", "t5x", "encoder-decoder", "en", "dataset:EleutherAI/pile", "arxiv:2101.00027", "arxiv:2201.07311", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2024-01-16T12:49:45Z	2024-04-17T03:50:20+00:00	40	28	--- datasets: - EleutherAI/pile language: - en pipeline_tag: text2text-generation tags: - t5x - encoder-decoder --- Pile-T5 XXL is an Encoder-Decoder model trained on [the Pile](https://pile.eleuther.ai/) using the [T5x](https://github.com/google-research/t5x) library. The model was trained for 2 million steps or roughly 2 trillion tokens using MLM-objective similar to the original T5 model. The HF version of Pile-T5 XXL borrows UMT5's model implementation as it uses scalable model implementation from T5x and uses `LlamaTokenizer`. ### Model Details - Developed by: [EleutherAI](http://eleuther.ai) - Model type: Transformer-based Language Model - Language: English - Learn more: [Blogpost](). For details about the training dataset, see [the Pile paper](https://arxiv.org/abs/2101.00027), and [its data sheet](https://arxiv.org/abs/2201.07311). - License: Apache 2.0 - Contact: to ask questions about this model, join the [EleutherAI Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`. Please read the existing GPT-NeoX-20B documentation before asking about the model on Discord. For general correspondence: [contact@eleuther. ai](mailto:[email protected]). <figure style="width:30em"> \| Hyperparameter \| Value \| \| -------------------------- \| ----------- \| \| n<sub>parameters</sub> \| 11135426560 \| \| n<sub>encoder layers</sub> \| 24 \| \| n<sub>decoder layers</sub> \| 24 \| \| d<sub>model</sub> \| 10240 \| \| d<sub>emb</sub> \| 4096 \| \| n<sub>heads</sub> \| 64 \| \| d<sub>head</sub> \| 64 \| \| n<sub>vocab</sub> \| 32128 \| \| Sequence Length \| 512 \| </figure> ### Uses and limitations #### Intended use Pile-T5 was developed primarily for research purposes. It learns an inner representation of the English language that can be used to extract features useful for downstream tasks. In addition to scientific uses, you may also further fine-tune and adapt Pile-T5 for deployment, as long as your use is in accordance with the Apache 2.0 license. This model works with the [Transformers Library](https://huggingface.co/docs/transformers/index). If you decide to use pre-trained Pile-T5 as a basis for your fine-tuned model, please note that you need to conduct your own risk and bias assessment. #### Out-of-scope use Pile-T5 is not intended for deployment as-is. It is not a product and cannot be used for human-facing interactions without supervision. Pile-T5 has not been fine-tuned for downstream tasks for which language models are commonly deployed, such as writing genre prose, or commercial chatbots. This means Pile-T5 will likely not respond to a given prompt the way products such as ChatGPT do. This is because, unlike Pile-T5, ChatGPT was fine-tuned using methods such as Reinforcement Learning from Human Feedback (RLHF) to better “understand” human instructions and dialogue. This model is English-language only, and thus cannot be used for translation or generating text in other languages. #### Limitations and biases The core functionality of Pile-T5 is to take a string of text that has been partially replaced with mask tokens and predict a sequence of tokens that would replace those mask tokens. Remember that the statistically most likely sequence of tokens need not result in the most “accurate” text. Never rely on Pile-T5 to produce factually accurate output. This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset known to contain profanity and texts that are lewd or otherwise offensive. See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a discussion of documented biases with regards to gender, religion, and race. Pile-T5 may produce socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. We recommend curating the outputs of this model before presenting it to a human reader. Please inform your audience that you are using artificially generated text. #### How to use Pile-T5 can be loaded using the `AutoModelForSeq2SeqLM` functionality: ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("EleutherAI/pile-t5-xxl") model = AutoModelForSeq2SeqLM.from_pretrained("EleutherAI/pile-t5-xxl") ``` ### Training #### Training dataset The Pile is a 825GiB general-purpose dataset in English. It was created by EleutherAI specifically for training large language models. It contains texts from 22 diverse sources, roughly broken down into five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl), prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and miscellaneous (e.g. GitHub, Enron Emails). See [the Pile paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources, methodology, and a discussion of ethical implications. Consult [the datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation about the Pile and its component datasets. The Pile can be downloaded from the [official website](https://pile.eleuther.ai/), or from a [community mirror](https://the-eye.eu/public/AI/pile/). The Pile was deduplicated before being used to train Pile-T5. #### Training procedure Pile-T5 was trained with a batch size of approximately 1M tokens (2048 sequences of 512 tokens each), for a total of 2,000,000 steps. Pile-T5 was trained with the span-corruption objective. #### Training checkpoints Intermediate checkpoints for Pile-T5 are accessible within this repository. There are in total 200 checkpoints that are spaced 10,000 steps. For T5x-native checkpoints that can be used for finetuning with the T5x library, refer to [here](https://huggingface.co/lintang/pile-t5-xxl-t5x) The training loss (in tfevent format) and validation perplexity (in jsonl) can be found [here](https://huggingface.co/EleutherAI/pile-t5-xxl/blob/main/xxl.zip). ### Evaluations Pile-T5 XXL was evaluated on SuperGLUE, CodeXGLUE. A Flan-finetuned version was evaluated on Flan Held In tasks, MMLU and BBH. Results can be seen in the [blogpost](https://blog.eleuther.ai/pile-t5/) ### BibTeX ``` @misc{2024PileT5, author = {Lintang Sutawika and Aran Komatsuzaki and Colin Raffel}, title = {Pile-T5}, year = {2024}, url = {https://blog.eleuther.ai/pile-t5/}, note = {Blog post}, } ```	null	Non_BioNLP	Pile-T5 XXL is an Encoder-Decoder model trained on [the Pile](https://pile.eleuther.ai/) using the [T5x](https://github.com/google-research/t5x) library. The model was trained for 2 million steps or roughly 2 trillion tokens using MLM-objective similar to the original T5 model. The HF version of Pile-T5 XXL borrows UMT5's model implementation as it uses scalable model implementation from T5x and uses `LlamaTokenizer`. ### Model Details - Developed by: [EleutherAI](http://eleuther.ai) - Model type: Transformer-based Language Model - Language: English - Learn more: [Blogpost](). For details about the training dataset, see [the Pile paper](https://arxiv.org/abs/2101.00027), and [its data sheet](https://arxiv.org/abs/2201.07311). - License: Apache 2.0 - Contact: to ask questions about this model, join the [EleutherAI Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`. Please read the existing GPT-NeoX-20B documentation before asking about the model on Discord. For general correspondence: [contact@eleuther. ai](mailto:[email protected]). <figure style="width:30em"> \| Hyperparameter \| Value \| \| -------------------------- \| ----------- \| \| n<sub>parameters</sub> \| 11135426560 \| \| n<sub>encoder layers</sub> \| 24 \| \| n<sub>decoder layers</sub> \| 24 \| \| d<sub>model</sub> \| 10240 \| \| d<sub>emb</sub> \| 4096 \| \| n<sub>heads</sub> \| 64 \| \| d<sub>head</sub> \| 64 \| \| n<sub>vocab</sub> \| 32128 \| \| Sequence Length \| 512 \| </figure> ### Uses and limitations #### Intended use Pile-T5 was developed primarily for research purposes. It learns an inner representation of the English language that can be used to extract features useful for downstream tasks. In addition to scientific uses, you may also further fine-tune and adapt Pile-T5 for deployment, as long as your use is in accordance with the Apache 2.0 license. This model works with the [Transformers Library](https://huggingface.co/docs/transformers/index). If you decide to use pre-trained Pile-T5 as a basis for your fine-tuned model, please note that you need to conduct your own risk and bias assessment. #### Out-of-scope use Pile-T5 is not intended for deployment as-is. It is not a product and cannot be used for human-facing interactions without supervision. Pile-T5 has not been fine-tuned for downstream tasks for which language models are commonly deployed, such as writing genre prose, or commercial chatbots. This means Pile-T5 will likely not respond to a given prompt the way products such as ChatGPT do. This is because, unlike Pile-T5, ChatGPT was fine-tuned using methods such as Reinforcement Learning from Human Feedback (RLHF) to better “understand” human instructions and dialogue. This model is English-language only, and thus cannot be used for translation or generating text in other languages. #### Limitations and biases The core functionality of Pile-T5 is to take a string of text that has been partially replaced with mask tokens and predict a sequence of tokens that would replace those mask tokens. Remember that the statistically most likely sequence of tokens need not result in the most “accurate” text. Never rely on Pile-T5 to produce factually accurate output. This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset known to contain profanity and texts that are lewd or otherwise offensive. See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a discussion of documented biases with regards to gender, religion, and race. Pile-T5 may produce socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. We recommend curating the outputs of this model before presenting it to a human reader. Please inform your audience that you are using artificially generated text. #### How to use Pile-T5 can be loaded using the `AutoModelForSeq2SeqLM` functionality: ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("EleutherAI/pile-t5-xxl") model = AutoModelForSeq2SeqLM.from_pretrained("EleutherAI/pile-t5-xxl") ``` ### Training #### Training dataset The Pile is a 825GiB general-purpose dataset in English. It was created by EleutherAI specifically for training large language models. It contains texts from 22 diverse sources, roughly broken down into five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl), prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and miscellaneous (e.g. GitHub, Enron Emails). See [the Pile paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources, methodology, and a discussion of ethical implications. Consult [the datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation about the Pile and its component datasets. The Pile can be downloaded from the [official website](https://pile.eleuther.ai/), or from a [community mirror](https://the-eye.eu/public/AI/pile/). The Pile was deduplicated before being used to train Pile-T5. #### Training procedure Pile-T5 was trained with a batch size of approximately 1M tokens (2048 sequences of 512 tokens each), for a total of 2,000,000 steps. Pile-T5 was trained with the span-corruption objective. #### Training checkpoints Intermediate checkpoints for Pile-T5 are accessible within this repository. There are in total 200 checkpoints that are spaced 10,000 steps. For T5x-native checkpoints that can be used for finetuning with the T5x library, refer to [here](https://huggingface.co/lintang/pile-t5-xxl-t5x) The training loss (in tfevent format) and validation perplexity (in jsonl) can be found [here](https://huggingface.co/EleutherAI/pile-t5-xxl/blob/main/xxl.zip). ### Evaluations Pile-T5 XXL was evaluated on SuperGLUE, CodeXGLUE. A Flan-finetuned version was evaluated on Flan Held In tasks, MMLU and BBH. Results can be seen in the [blogpost](https://blog.eleuther.ai/pile-t5/) ### BibTeX ``` @misc{2024PileT5, author = {Lintang Sutawika and Aran Komatsuzaki and Colin Raffel}, title = {Pile-T5}, year = {2024}, url = {https://blog.eleuther.ai/pile-t5/}, note = {Blog post}, } ```	{"datasets": ["EleutherAI/pile"], "language": ["en"], "pipeline_tag": "text2text-generation", "tags": ["t5x", "encoder-decoder"]}	task	[ "TRANSLATION" ]	44,580
VityaVitalich/bert-base-cased-sst2	VityaVitalich	text-classification	[ "transformers", "pytorch", "bert", "text-classification", "generated_from_trainer", "dataset:sst2", "base_model:google-bert/bert-base-cased", "base_model:finetune:google-bert/bert-base-cased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-10-02T14:39:42Z	2023-10-02T14:40:10+00:00	9	0	--- base_model: bert-base-cased datasets: - sst2 license: apache-2.0 metrics: - accuracy tags: - generated_from_trainer model-index: - name: bert-base-cased-sst2 results: - task: type: text-classification name: Text Classification dataset: name: sst2 type: sst2 config: default split: validation args: default metrics: - type: accuracy value: 0.9139908256880734 name: Accuracy --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-cased-sst2 This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the sst2 dataset. It achieves the following results on the evaluation set: - Loss: 0.2103 - Accuracy: 0.9140 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 128 - eval_batch_size: 128 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.2267 \| 1.0 \| 527 \| 0.2103 \| 0.9140 \| \| 0.1091 \| 2.0 \| 1054 \| 0.2637 \| 0.9174 \| \| 0.0722 \| 3.0 \| 1581 \| 0.2673 \| 0.9174 \| \| 0.0467 \| 4.0 \| 2108 \| 0.2947 \| 0.9266 \| \| 0.0298 \| 5.0 \| 2635 \| 0.3344 \| 0.9209 \| ### Framework versions - Transformers 4.34.0.dev0 - Pytorch 2.0.1+cu117 - Datasets 2.14.5 - Tokenizers 0.14.0	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-cased-sst2 This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the sst2 dataset. It achieves the following results on the evaluation set: - Loss: 0.2103 - Accuracy: 0.9140 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 128 - eval_batch_size: 128 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.2267 \| 1.0 \| 527 \| 0.2103 \| 0.9140 \| \| 0.1091 \| 2.0 \| 1054 \| 0.2637 \| 0.9174 \| \| 0.0722 \| 3.0 \| 1581 \| 0.2673 \| 0.9174 \| \| 0.0467 \| 4.0 \| 2108 \| 0.2947 \| 0.9266 \| \| 0.0298 \| 5.0 \| 2635 \| 0.3344 \| 0.9209 \| ### Framework versions - Transformers 4.34.0.dev0 - Pytorch 2.0.1+cu117 - Datasets 2.14.5 - Tokenizers 0.14.0	{"base_model": "bert-base-cased", "datasets": ["sst2"], "license": "apache-2.0", "metrics": ["accuracy"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-cased-sst2", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "sst2", "type": "sst2", "config": "default", "split": "validation", "args": "default"}, "metrics": [{"type": "accuracy", "value": 0.9139908256880734, "name": "Accuracy"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,581
dreamboat26/NLP_Transformer_1	dreamboat26	translation	[ "translation", "dataset:Helsinki-NLP/tatoeba_mt", "license:openrail", "region:us" ]	2023-08-11T19:27:01Z	2023-08-11T19:32:23+00:00	0	0	--- datasets: - Helsinki-NLP/tatoeba_mt license: openrail pipeline_tag: translation --- This is the basic implementation of what Transformers do and a few examples to show how to implement it	null	Non_BioNLP	This is the basic implementation of what Transformers do and a few examples to show how to implement it	{"datasets": ["Helsinki-NLP/tatoeba_mt"], "license": "openrail", "pipeline_tag": "translation"}	task	[ "TRANSLATION" ]	44,582
mtspeech/MooER-MTL-5K	mtspeech	automatic-speech-recognition	[ "asr", "automatic-speech-recognition", "automatic-speech-translation", "speech-translation", "speech-recognition", "zh", "en", "arxiv:2408.05101", "license:mit", "region:us" ]	2024-08-09T04:50:51Z	2024-08-27T12:41:18+00:00	0	5	--- language: - zh - en license: mit metrics: - cer - bleu tags: - asr - automatic-speech-recognition - automatic-speech-translation - speech-translation - speech-recognition --- # MooER (摩耳): an LLM-based Speech Recognition and Translation Model from Moore Threads Online Demo: https://mooer-speech.mthreads.com:10077/ ## 🔥 Update We release a new model MooER-80K-v2 using 80K hours of data. Click [here](https://huggingface.co/mtspeech/MooER-MTL-80K) to try the new model. ## 📖 Introduction We introduce MooER (摩耳): an LLM-based speech recognition and translation model developed by Moore Threads. With the MooER framework, you can transcribe the speech into text (speech recognition or, ASR), and translate it into other languages (speech translation or, AST) in a end-to-end manner. The performance of MooER is demonstrated in the subsequent section, along with our insights into model configurations, training strategies, and more, provided in our [technical report](https://arxiv.org/abs/2408.05101). For the usage of the model files, please refer to our [GitHub](https://github.com/MooreThreads/MooER) <br> <p align="center"> <img src="assets/framework.png" width="600"/> <p> <br> ## 🥊 Evaluation Results We demonstrate the training data and the evaluation results below. For more comprehensive information, please refer to our [report](https://arxiv.org/pdf/2408.05101). ### Training data We utilize 5k hours of data (MT5K) to train our basic MooER-5K model. The data sources include: \| Dataset \| Duration \| \|---------------\|---------------\| \| aishell2 \| 137h \| \| librispeech \| 131h \| \| multi_cn \| 100h \| \| wenetspeech \| 1361h \| \| in-house data \| 3274h \| Note that, data from the open-source datasets were randomly selected from the full training set. The in-house data, collected internally without text, were transcribed using a third-party ASR service. Since all the above datasets were originally designed only for the speech recognition task, no translation results are available. To train our speech translation model, we used a third-party translation service to generate pseudo-labels. No data filtering techniques were applied. At this moment, we are also developing a new model trained with 80K hours of data. ### Speech Recognition The performance of speech recognition is evaluated using WER/CER. <table> <tr> <th>Language</th> <th>Testset</th> <th>Paraformer-large</th> <th>SenseVoice-small</th> <th>Qwen-audio</th> <th>Whisper-large-v3</th> <th>SeamlessM4T-v2</th> <th>MooER-5K</th> <th>MooER-80K</th> <th>MooER-80K-v2</th> </tr> <tr> <td rowspan="7">Chinese</td> <td>aishell1</td> <td>1.93</td> <td>3.03</td> <td>1.43</td> <td>7.86</td> <td>4.09</td> <td>1.93</td> <td>1.25</td> <td>1.00</td> </tr> <tr> <td>aishell2_ios</td> <td>2.85</td> <td>3.79</td> <td>3.57</td> <td>5.38</td> <td>4.81</td> <td>3.17</td> <td>2.67</td> <td>2.62</td> </tr> <tr> <td>test_magicdata</td> <td>3.66</td> <td>3.81</td> <td>5.31</td> <td>8.36</td> <td>9.69</td> <td>3.48</td> <td>2.52</td> <td>2.17</td> </tr> <tr> <td>test_thchs</td> <td>3.99</td> <td>5.17</td> <td>4.86</td> <td>9.06</td> <td>7.14</td> <td>4.11</td> <td>3.14</td> <td>3.00</td> </tr> <tr> <td>fleurs cmn_dev</td> <td>5.56</td> <td>6.39</td> <td>10.54</td> <td>4.54</td> <td>7.12</td> <td>5.81</td> <td>5.23</td> <td>5.15</td> </tr> <tr> <td>fleurs cmn_test</td> <td>6.92</td> <td>7.36</td> <td>11.07</td> <td>5.24</td> <td>7.66</td> <td>6.77</td> <td>6.18</td> <td>6.14</td> </tr> <tr> <td>average</td> <td><strong>4.15</strong></td> <td><strong>4.93</strong></td> <td><strong>6.13</strong></td> <td><strong>6.74</strong></td> <td><strong>6.75</strong></td> <td><strong>4.21</strong></td> <td><strong>3.50</strong></td> <td><strong>3.35</strong></td> </tr> <tr> <td rowspan="7">English</td> <td>librispeech test_clean</td> <td>14.15</td> <td>4.07</td> <td>2.15</td> <td>3.42</td> <td>2.77</td> <td>7.78</td> <td>4.11</td> <td>3.57</td> </tr> <tr> <td>librispeech test_other</td> <td>22.99</td> <td>8.26</td> <td>4.68</td> <td>5.62</td> <td>5.25</td> <td>15.25</td> <td>9.99</td> <td>9.09</td> </tr> <tr> <td>fleurs eng_dev</td> <td>24.93</td> <td>12.92</td> <td>22.53</td> <td>11.63</td> <td>11.36</td> <td>18.89</td> <td>13.32</td> <td>13.12</td> </tr> <tr> <td>fleurs eng_test</td> <td>26.81</td> <td>13.41</td> <td>22.51</td> <td>12.57</td> <td>11.82</td> <td>20.41</td> <td>14.97</td> <td>14.74</td> </tr> <tr> <td>gigaspeech dev</td> <td>24.23</td> <td>19.44</td> <td>12.96</td> <td>19.18</td> <td>28.01</td> <td>23.46</td> <td>16.92</td> <td>17.34</td> </tr> <tr> <td>gigaspeech test</td> <td>23.07</td> <td>16.65</td> <td>13.26</td> <td>22.34</td> <td>28.65</td> <td>22.09</td> <td>16.64</td> <td>16.97</td> </tr> <tr> <td>average</td> <td><strong>22.70</strong></td> <td><strong>12.46</strong></td> <td><strong>13.02</strong></td> <td><strong>12.46</strong></td> <td><strong>14.64</strong></td> <td><strong>17.98</strong></td> <td><strong>12.66</strong></td> <td><strong>12.47</strong></td> </tr> </table> ### Speech Translation (zh -> en) For speech translation, the performanced is evaluated using BLEU score. \| Testset \| Speech-LLaMA \| Whisper-large-v3 \| Qwen-audio \| Qwen2-audio \| SeamlessM4T-v2 \| MooER-5K \| MooER-5K-MTL \| \|--------\|-------------\|-------------------\|------------\|-------------\|-----------------\|--------\|--------------\| \|CoVoST1 zh2en \| - \| 13.5 \| 13.5 \| - \| 25.3 \| - \| 30.2 \| \|CoVoST2 zh2en \| 12.3 \| 12.2 \| 15.7 \| 24.4 \| 22.2 \| 23.4 \| 25.2 \| \|CCMT2019 dev \| - \| 15.9 \| 12.0 \| - \| 14.8 \| - \| 19.6 \| ## 🏁 Getting Started Please visit our [GitHub](https://github.com/MooreThreads/MooER) for the setup and usage. ## 🧾 License Please see the [LICENSE](LICENSE). ## 💖 Citation If you find MooER useful for your research, please 🌟 this repo and cite our work using the following BibTeX: ```bibtex @article{liang2024mooer, title = {MooER: an LLM-based Speech Recognition and Translation Model from Moore Threads}, author = {Zhenlin Liang, Junhao Xu, Yi Liu, Yichao Hu, Jian Li, Yajun Zheng, Meng Cai, Hua Wang}, journal = {arXiv preprint arXiv:2408.05101}, url = {https://arxiv.org/abs/2408.05101}, year = {2024} } ``` ## 📧 Contact If you encouter any problems, feel free to create a discussion. Moore Threads Website: https://www.mthreads.com/ <br> <p align="left"> <img src="assets/MTLogo.png" width="300"/> <p> <br>	null	Non_BioNLP	# MooER (摩耳): an LLM-based Speech Recognition and Translation Model from Moore Threads Online Demo: https://mooer-speech.mthreads.com:10077/ ## 🔥 Update We release a new model MooER-80K-v2 using 80K hours of data. Click [here](https://huggingface.co/mtspeech/MooER-MTL-80K) to try the new model. ## 📖 Introduction We introduce MooER (摩耳): an LLM-based speech recognition and translation model developed by Moore Threads. With the MooER framework, you can transcribe the speech into text (speech recognition or, ASR), and translate it into other languages (speech translation or, AST) in a end-to-end manner. The performance of MooER is demonstrated in the subsequent section, along with our insights into model configurations, training strategies, and more, provided in our [technical report](https://arxiv.org/abs/2408.05101). For the usage of the model files, please refer to our [GitHub](https://github.com/MooreThreads/MooER) <br> <p align="center"> <img src="assets/framework.png" width="600"/> <p> <br> ## 🥊 Evaluation Results We demonstrate the training data and the evaluation results below. For more comprehensive information, please refer to our [report](https://arxiv.org/pdf/2408.05101). ### Training data We utilize 5k hours of data (MT5K) to train our basic MooER-5K model. The data sources include: \| Dataset \| Duration \| \|---------------\|---------------\| \| aishell2 \| 137h \| \| librispeech \| 131h \| \| multi_cn \| 100h \| \| wenetspeech \| 1361h \| \| in-house data \| 3274h \| Note that, data from the open-source datasets were randomly selected from the full training set. The in-house data, collected internally without text, were transcribed using a third-party ASR service. Since all the above datasets were originally designed only for the speech recognition task, no translation results are available. To train our speech translation model, we used a third-party translation service to generate pseudo-labels. No data filtering techniques were applied. At this moment, we are also developing a new model trained with 80K hours of data. ### Speech Recognition The performance of speech recognition is evaluated using WER/CER. <table> <tr> <th>Language</th> <th>Testset</th> <th>Paraformer-large</th> <th>SenseVoice-small</th> <th>Qwen-audio</th> <th>Whisper-large-v3</th> <th>SeamlessM4T-v2</th> <th>MooER-5K</th> <th>MooER-80K</th> <th>MooER-80K-v2</th> </tr> <tr> <td rowspan="7">Chinese</td> <td>aishell1</td> <td>1.93</td> <td>3.03</td> <td>1.43</td> <td>7.86</td> <td>4.09</td> <td>1.93</td> <td>1.25</td> <td>1.00</td> </tr> <tr> <td>aishell2_ios</td> <td>2.85</td> <td>3.79</td> <td>3.57</td> <td>5.38</td> <td>4.81</td> <td>3.17</td> <td>2.67</td> <td>2.62</td> </tr> <tr> <td>test_magicdata</td> <td>3.66</td> <td>3.81</td> <td>5.31</td> <td>8.36</td> <td>9.69</td> <td>3.48</td> <td>2.52</td> <td>2.17</td> </tr> <tr> <td>test_thchs</td> <td>3.99</td> <td>5.17</td> <td>4.86</td> <td>9.06</td> <td>7.14</td> <td>4.11</td> <td>3.14</td> <td>3.00</td> </tr> <tr> <td>fleurs cmn_dev</td> <td>5.56</td> <td>6.39</td> <td>10.54</td> <td>4.54</td> <td>7.12</td> <td>5.81</td> <td>5.23</td> <td>5.15</td> </tr> <tr> <td>fleurs cmn_test</td> <td>6.92</td> <td>7.36</td> <td>11.07</td> <td>5.24</td> <td>7.66</td> <td>6.77</td> <td>6.18</td> <td>6.14</td> </tr> <tr> <td>average</td> <td><strong>4.15</strong></td> <td><strong>4.93</strong></td> <td><strong>6.13</strong></td> <td><strong>6.74</strong></td> <td><strong>6.75</strong></td> <td><strong>4.21</strong></td> <td><strong>3.50</strong></td> <td><strong>3.35</strong></td> </tr> <tr> <td rowspan="7">English</td> <td>librispeech test_clean</td> <td>14.15</td> <td>4.07</td> <td>2.15</td> <td>3.42</td> <td>2.77</td> <td>7.78</td> <td>4.11</td> <td>3.57</td> </tr> <tr> <td>librispeech test_other</td> <td>22.99</td> <td>8.26</td> <td>4.68</td> <td>5.62</td> <td>5.25</td> <td>15.25</td> <td>9.99</td> <td>9.09</td> </tr> <tr> <td>fleurs eng_dev</td> <td>24.93</td> <td>12.92</td> <td>22.53</td> <td>11.63</td> <td>11.36</td> <td>18.89</td> <td>13.32</td> <td>13.12</td> </tr> <tr> <td>fleurs eng_test</td> <td>26.81</td> <td>13.41</td> <td>22.51</td> <td>12.57</td> <td>11.82</td> <td>20.41</td> <td>14.97</td> <td>14.74</td> </tr> <tr> <td>gigaspeech dev</td> <td>24.23</td> <td>19.44</td> <td>12.96</td> <td>19.18</td> <td>28.01</td> <td>23.46</td> <td>16.92</td> <td>17.34</td> </tr> <tr> <td>gigaspeech test</td> <td>23.07</td> <td>16.65</td> <td>13.26</td> <td>22.34</td> <td>28.65</td> <td>22.09</td> <td>16.64</td> <td>16.97</td> </tr> <tr> <td>average</td> <td><strong>22.70</strong></td> <td><strong>12.46</strong></td> <td><strong>13.02</strong></td> <td><strong>12.46</strong></td> <td><strong>14.64</strong></td> <td><strong>17.98</strong></td> <td><strong>12.66</strong></td> <td><strong>12.47</strong></td> </tr> </table> ### Speech Translation (zh -> en) For speech translation, the performanced is evaluated using BLEU score. \| Testset \| Speech-LLaMA \| Whisper-large-v3 \| Qwen-audio \| Qwen2-audio \| SeamlessM4T-v2 \| MooER-5K \| MooER-5K-MTL \| \|--------\|-------------\|-------------------\|------------\|-------------\|-----------------\|--------\|--------------\| \|CoVoST1 zh2en \| - \| 13.5 \| 13.5 \| - \| 25.3 \| - \| 30.2 \| \|CoVoST2 zh2en \| 12.3 \| 12.2 \| 15.7 \| 24.4 \| 22.2 \| 23.4 \| 25.2 \| \|CCMT2019 dev \| - \| 15.9 \| 12.0 \| - \| 14.8 \| - \| 19.6 \| ## 🏁 Getting Started Please visit our [GitHub](https://github.com/MooreThreads/MooER) for the setup and usage. ## 🧾 License Please see the [LICENSE](LICENSE). ## 💖 Citation If you find MooER useful for your research, please 🌟 this repo and cite our work using the following BibTeX: ```bibtex @article{liang2024mooer, title = {MooER: an LLM-based Speech Recognition and Translation Model from Moore Threads}, author = {Zhenlin Liang, Junhao Xu, Yi Liu, Yichao Hu, Jian Li, Yajun Zheng, Meng Cai, Hua Wang}, journal = {arXiv preprint arXiv:2408.05101}, url = {https://arxiv.org/abs/2408.05101}, year = {2024} } ``` ## 📧 Contact If you encouter any problems, feel free to create a discussion. Moore Threads Website: https://www.mthreads.com/ <br> <p align="left"> <img src="assets/MTLogo.png" width="300"/> <p> <br>	{"language": ["zh", "en"], "license": "mit", "metrics": ["cer", "bleu"], "tags": ["asr", "automatic-speech-recognition", "automatic-speech-translation", "speech-translation", "speech-recognition"]}	task	[ "TRANSLATION" ]	44,583
machinelearningzuu/paper-summarization	machinelearningzuu	text2text-generation	[ "transformers", "pytorch", "tensorboard", "t5", "text2text-generation", "generated_from_trainer", "base_model:google-t5/t5-small", "base_model:finetune:google-t5/t5-small", "license:apache-2.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2023-07-13T14:31:12Z	2023-07-29T01:16:54+00:00	14	0	--- base_model: t5-small license: apache-2.0 tags: - generated_from_trainer model-index: - name: paper-summarization results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # paper-summarization This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.3296 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:----:\|:---------------:\| \| 3.2336 \| 1.0 \| 78 \| 2.5990 \| \| 2.7888 \| 2.0 \| 156 \| 2.3754 \| \| 2.5667 \| 3.0 \| 234 \| 2.3296 \| ### Framework versions - Transformers 4.31.0 - Pytorch 1.13.1 - Datasets 2.12.0 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # paper-summarization This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.3296 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:----:\|:---------------:\| \| 3.2336 \| 1.0 \| 78 \| 2.5990 \| \| 2.7888 \| 2.0 \| 156 \| 2.3754 \| \| 2.5667 \| 3.0 \| 234 \| 2.3296 \| ### Framework versions - Transformers 4.31.0 - Pytorch 1.13.1 - Datasets 2.12.0 - Tokenizers 0.13.3	{"base_model": "t5-small", "license": "apache-2.0", "tags": ["generated_from_trainer"], "model-index": [{"name": "paper-summarization", "results": []}]}	task	[ "SUMMARIZATION" ]	44,584
nikolab/speecht5_tts_hr	nikolab	text-to-audio	[ "transformers", "safetensors", "speecht5", "text-to-audio", "hr", "dataset:facebook/voxpopuli", "arxiv:2110.07205", "base_model:microsoft/speecht5_tts", "base_model:finetune:microsoft/speecht5_tts", "license:mit", "endpoints_compatible", "region:us" ]	2024-09-30T14:07:27Z	2024-10-07T12:10:19+00:00	78	1	--- base_model: - microsoft/speecht5_tts datasets: - facebook/voxpopuli language: - hr library_name: transformers license: mit --- # speecht5_tts_hr This is a fine-tuned version of SpeechT5 text-to-speech model tailored for Croatian language. # Model SpeechT5 is a fine-tuned model for speech synthesis (text-to-speech) on the LibriTTS dataset. It was created as an upgraded version of the successful T5 model (Text-To-Text Transfer Transformer), which was trained only for natural language processing. The model was originally presented by the Microsoft research group, in the scientific paper "SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing" (https://arxiv.org/abs/2110.07205). The SpeechT5 model was also chosen due to the extensive evaluation, carried out in the mentioned scientific work, which showed very good results in a wide range of speech processing tasks, including automatic speech recognition, speech synthesis, speech translation, voice conversion, speech enhancement and speaker identification. SpeechT5 contains three types of speech models in one architecture. The model can be used to convert: - speech to text - for automatic speech recognition or speaker identification, - text to speech - for sound synthesis, - speech to speech - to convert between different voices or improve speech. The SpeechT5 model consists of a common network of encoders and decoders, with an additional six neural networks that are specific to the particular modality of the data being processed (speech/text). The unique thing about the SpeechT5 model is that the model is first pre-trained on different speech-to-text and text-to-speech data modalities, so that it learns in the unified representation space of both text and speech. In this way, the model learns from text and speech at the same time. This allows us to fine-tune the pre-trained model for different tasks, such as text-to-speech, in ex-yu languages (Montenegrin, Serbian, Bosnian, Croatian). # Dataset LibriTTS (https://www.openslr.org/60/) is a multi-speaker English corpus of approximately 585 hours of English speech, prepared by Heiga Zen with the help of members of the Google Speech and Google Brain teams. This corpus is designed for TTS (text-to-speech) research. It is derived from the original LibriSpeech corpus (https://www.openslr.org/12/) - mp3 audio files from LibriVox and text files from Project Gutenberg. The VoxPopuli dataset, published in the scientific paper link, contains: - 400 thousand hours of untagged voice data for 23 languages, - 1.8 thousand hours of transcribed speech data for 16 languages, - 17.3 thousand hours of "speech-to-speech" data, - 29 hours of transcribed speech data of non-native English speakers, intended for research into accented speech. # Technical implementation Experimental trainings of the SpeechT5 model were carried out with the aim of adapting the basic model for the use of text-to-speech conversion. As the original SpeechT5 model was trained on tasks exclusively in English (LibriTTS dataset), it was necessary to implement the training of the new model, on the available data in Croatian language. One of the popular open datasets for this use is the VoxPopuli set, which contains sound recordings of the European Parliament from 2009 to 2020. Given that data in all regional languages is not available to the required extent, data in the Croatian language, which is the most represented, was taken from the VoxPopuli dataset. In the next stages of the project, data will be collected in Montenegrin, Serbian and Bosnian languages, in order to improve the quality of training and the accuracy of the model. Thus, the final dataset consists of 43 transcribed hours of speech, 83 different speakers and 337 thousand transcribed tokens (1 token = 3/4 words). In the first phase of technical implementation, the dataset went through several stages of processing in order to adapt and standardize it for training the SpeechT5 model. Data processing methods belong to the standard methods of linguistic data manipulation in the field of natural language processing (vocabulary formation, tokenization, removal or conversion of unsupported characters/letters, text/speech cleaning, text normalization). In the next phase, the statistics of speakers in the VoxPopuli dataset were analyzed, based on which speakers with satisfactory text/speech quality and a sufficient number of samples for model training were selected. In this phase, the balancing of the dataset was carried out so that both male and female speakers, with high-quality text/speech samples, were equally represented in the training. After the preparation of the data, the adjustment and optimization of the hyperparameters of the SpeechT5 model, which are necessary so that the training of the model can be performed quickly and efficiently, with satisfactory accuracy, was started. Several experimental training sessions were performed to obtain the optimal hyperparameters, which were then used in the evaluation phase of the model. The evaluation of the model, on the dataset intended for testing, showed promising results. The model started to learn on the prepared dataset, but it also showed certain limitations. The main limitation is related to the length of the input text sequence. The model showed the inability to generate speech for long sequences of input text (over 20 words). The limitation was overcome by dividing the input sequence into smaller units and in that form passed to the model for processing. The main reason for the appearance of this limitation lies primarily in the lack of a large amount of data on which it is necessary to fine-tune the model in order to obtain the best possible results.	null	Non_BioNLP	# speecht5_tts_hr This is a fine-tuned version of SpeechT5 text-to-speech model tailored for Croatian language. # Model SpeechT5 is a fine-tuned model for speech synthesis (text-to-speech) on the LibriTTS dataset. It was created as an upgraded version of the successful T5 model (Text-To-Text Transfer Transformer), which was trained only for natural language processing. The model was originally presented by the Microsoft research group, in the scientific paper "SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing" (https://arxiv.org/abs/2110.07205). The SpeechT5 model was also chosen due to the extensive evaluation, carried out in the mentioned scientific work, which showed very good results in a wide range of speech processing tasks, including automatic speech recognition, speech synthesis, speech translation, voice conversion, speech enhancement and speaker identification. SpeechT5 contains three types of speech models in one architecture. The model can be used to convert: - speech to text - for automatic speech recognition or speaker identification, - text to speech - for sound synthesis, - speech to speech - to convert between different voices or improve speech. The SpeechT5 model consists of a common network of encoders and decoders, with an additional six neural networks that are specific to the particular modality of the data being processed (speech/text). The unique thing about the SpeechT5 model is that the model is first pre-trained on different speech-to-text and text-to-speech data modalities, so that it learns in the unified representation space of both text and speech. In this way, the model learns from text and speech at the same time. This allows us to fine-tune the pre-trained model for different tasks, such as text-to-speech, in ex-yu languages (Montenegrin, Serbian, Bosnian, Croatian). # Dataset LibriTTS (https://www.openslr.org/60/) is a multi-speaker English corpus of approximately 585 hours of English speech, prepared by Heiga Zen with the help of members of the Google Speech and Google Brain teams. This corpus is designed for TTS (text-to-speech) research. It is derived from the original LibriSpeech corpus (https://www.openslr.org/12/) - mp3 audio files from LibriVox and text files from Project Gutenberg. The VoxPopuli dataset, published in the scientific paper link, contains: - 400 thousand hours of untagged voice data for 23 languages, - 1.8 thousand hours of transcribed speech data for 16 languages, - 17.3 thousand hours of "speech-to-speech" data, - 29 hours of transcribed speech data of non-native English speakers, intended for research into accented speech. # Technical implementation Experimental trainings of the SpeechT5 model were carried out with the aim of adapting the basic model for the use of text-to-speech conversion. As the original SpeechT5 model was trained on tasks exclusively in English (LibriTTS dataset), it was necessary to implement the training of the new model, on the available data in Croatian language. One of the popular open datasets for this use is the VoxPopuli set, which contains sound recordings of the European Parliament from 2009 to 2020. Given that data in all regional languages is not available to the required extent, data in the Croatian language, which is the most represented, was taken from the VoxPopuli dataset. In the next stages of the project, data will be collected in Montenegrin, Serbian and Bosnian languages, in order to improve the quality of training and the accuracy of the model. Thus, the final dataset consists of 43 transcribed hours of speech, 83 different speakers and 337 thousand transcribed tokens (1 token = 3/4 words). In the first phase of technical implementation, the dataset went through several stages of processing in order to adapt and standardize it for training the SpeechT5 model. Data processing methods belong to the standard methods of linguistic data manipulation in the field of natural language processing (vocabulary formation, tokenization, removal or conversion of unsupported characters/letters, text/speech cleaning, text normalization). In the next phase, the statistics of speakers in the VoxPopuli dataset were analyzed, based on which speakers with satisfactory text/speech quality and a sufficient number of samples for model training were selected. In this phase, the balancing of the dataset was carried out so that both male and female speakers, with high-quality text/speech samples, were equally represented in the training. After the preparation of the data, the adjustment and optimization of the hyperparameters of the SpeechT5 model, which are necessary so that the training of the model can be performed quickly and efficiently, with satisfactory accuracy, was started. Several experimental training sessions were performed to obtain the optimal hyperparameters, which were then used in the evaluation phase of the model. The evaluation of the model, on the dataset intended for testing, showed promising results. The model started to learn on the prepared dataset, but it also showed certain limitations. The main limitation is related to the length of the input text sequence. The model showed the inability to generate speech for long sequences of input text (over 20 words). The limitation was overcome by dividing the input sequence into smaller units and in that form passed to the model for processing. The main reason for the appearance of this limitation lies primarily in the lack of a large amount of data on which it is necessary to fine-tune the model in order to obtain the best possible results.	{"base_model": ["microsoft/speecht5_tts"], "datasets": ["facebook/voxpopuli"], "language": ["hr"], "library_name": "transformers", "license": "mit"}	task	[ "TRANSLATION" ]	44,585
elaunlu/bert-base-uncased-finetuned-cola	elaunlu	text-classification	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-05-02T15:02:01Z	2023-05-04T18:27:57+00:00	12	0	--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: bert-base-uncased-finetuned-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: validation args: cola metrics: - type: matthews_correlation value: 0.518818601771926 name: Matthews Correlation --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4610 - Matthews Correlation: 0.5188 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.4985 \| 1.0 \| 535 \| 0.4610 \| 0.5188 \| ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4610 - Matthews Correlation: 0.5188 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Matthews Correlation \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------------:\| \| 0.4985 \| 1.0 \| 535 \| 0.4610 \| 0.5188 \| ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3	{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-uncased-finetuned-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "validation", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.518818601771926, "name": "Matthews Correlation"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,586
SEBIS/legal_t5_small_trans_cs_it_small_finetuned	SEBIS	text2text-generation	[ "transformers", "pytorch", "jax", "t5", "text2text-generation", "translation Cszech Italian model", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:04Z	2021-06-23T11:35:39+00:00	185	0	--- datasets: - dcep europarl jrc-acquis language: Cszech Italian tags: - translation Cszech Italian model widget: - text: Členové přítomní při závěrečném hlasování --- # legal_t5_small_trans_cs_it_small_finetuned model Model on translating legal text from Cszech to Italian. It was first released in [this repository](https://github.com/agemagician/LegalTrans). This model is first pretrained all the translation data over some unsupervised task. Then the model is trained on three parallel corpus from jrc-acquis, europarl and dcep. ## Model description legal_t5_small_trans_cs_it_small_finetuned is initially pretrained on unsupervised task with the all of the data of the training set. The unsupervised task was "masked language modelling". legal_t5_small_trans_cs_it_small_finetuned is based on the `t5-small` model and was trained on a large corpus of parallel text. This is a smaller model, which scales the baseline model of t5 down by using `dmodel = 512`, `dff = 2,048`, 8-headed attention, and only 6 layers each in the encoder and decoder. This variant has about 60 million parameters. ## Intended uses & limitations The model could be used for translation of legal texts from Cszech to Italian. ### How to use Here is how to use this model to translate legal text from Cszech to Italian in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_cs_it_small_finetuned"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_cs_it", do_lower_case=False, skip_special_tokens=True), device=0 ) cs_text = "Členové přítomní při závěrečném hlasování" pipeline([cs_text], max_length=512) ``` ## Training data The legal_t5_small_trans_cs_it_small_finetuned (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 5 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining The pre-training data was the combined data from all the 42 language pairs. The task for the model was to predict the portions of a sentence which were masked randomly. ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : \| Model \| BLEU score \| \|:-----:\|:-----:\| \| legal_t5_small_trans_cs_it_small_finetuned \| 46.367\| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) \| [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)	null	Non_BioNLP	# legal_t5_small_trans_cs_it_small_finetuned model Model on translating legal text from Cszech to Italian. It was first released in [this repository](https://github.com/agemagician/LegalTrans). This model is first pretrained all the translation data over some unsupervised task. Then the model is trained on three parallel corpus from jrc-acquis, europarl and dcep. ## Model description legal_t5_small_trans_cs_it_small_finetuned is initially pretrained on unsupervised task with the all of the data of the training set. The unsupervised task was "masked language modelling". legal_t5_small_trans_cs_it_small_finetuned is based on the `t5-small` model and was trained on a large corpus of parallel text. This is a smaller model, which scales the baseline model of t5 down by using `dmodel = 512`, `dff = 2,048`, 8-headed attention, and only 6 layers each in the encoder and decoder. This variant has about 60 million parameters. ## Intended uses & limitations The model could be used for translation of legal texts from Cszech to Italian. ### How to use Here is how to use this model to translate legal text from Cszech to Italian in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_cs_it_small_finetuned"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_cs_it", do_lower_case=False, skip_special_tokens=True), device=0 ) cs_text = "Členové přítomní při závěrečném hlasování" pipeline([cs_text], max_length=512) ``` ## Training data The legal_t5_small_trans_cs_it_small_finetuned (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 5 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining The pre-training data was the combined data from all the 42 language pairs. The task for the model was to predict the portions of a sentence which were masked randomly. ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : \| Model \| BLEU score \| \|:-----:\|:-----:\| \| legal_t5_small_trans_cs_it_small_finetuned \| 46.367\| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) \| [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)	{"datasets": ["dcep europarl jrc-acquis"], "language": "Cszech Italian", "tags": ["translation Cszech Italian model"], "widget": [{"text": "Členové přítomní při závěrečném hlasování"}]}	task	[ "TRANSLATION" ]	44,587
TalDeshe/autotrain-en-he-translation-89957144053	TalDeshe	translation	[ "transformers", "pytorch", "safetensors", "marian", "text2text-generation", "autotrain", "translation", "unk", "dataset:TalDeshe/autotrain-data-en-he-translation", "co2_eq_emissions", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-09-18T15:17:20Z	2023-09-18T15:24:41+00:00	21	1	--- datasets: - TalDeshe/autotrain-data-en-he-translation language: - unk - unk tags: - autotrain - translation co2_eq_emissions: emissions: 0.07162844137344286 --- # Model Trained Using AutoTrain - Problem type: Translation - Model ID: 89957144053 - CO2 Emissions (in grams): 0.0716 ## Validation Metrics - Loss: 0.370 - SacreBLEU: 75.619 - Gen len: 24.084	null	Non_BioNLP	# Model Trained Using AutoTrain - Problem type: Translation - Model ID: 89957144053 - CO2 Emissions (in grams): 0.0716 ## Validation Metrics - Loss: 0.370 - SacreBLEU: 75.619 - Gen len: 24.084	{"datasets": ["TalDeshe/autotrain-data-en-he-translation"], "language": ["unk", "unk"], "tags": ["autotrain", "translation"], "co2_eq_emissions": {"emissions": 0.07162844137344286}}	task	[ "TRANSLATION" ]	44,588
Izzet/qa_ytu_bert-base-turkish	Izzet	question-answering	[ "transformers", "pytorch", "bert", "question-answering", "license:mit", "endpoints_compatible", "region:us" ]	2022-09-03T23:19:56Z	2022-09-03T23:34:11+00:00	14	0	--- license: mit widget: - text: Ankara'da korumaya alınmış alanlar var mıdır? context: Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır. example_title: Ankara 1 - text: Ankara toprakları nelerden oluşur? context: Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır. example_title: Ankara 2 --- # Question Answering Model Fine-Tuned with YTU Dataset You can find detailed explanation about dataset [here](https://github.com/izzetkalic/botcuk-dataset-analyze/tree/main/datasets/qa-ytu).	null	Non_BioNLP	# Question Answering Model Fine-Tuned with YTU Dataset You can find detailed explanation about dataset [here](https://github.com/izzetkalic/botcuk-dataset-analyze/tree/main/datasets/qa-ytu).	{"license": "mit", "widget": [{"text": "Ankara'da korumaya alınmış alanlar var mıdır?", "context": "Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır.", "example_title": "Ankara 1"}, {"text": "Ankara toprakları nelerden oluşur?", "context": "Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır.", "example_title": "Ankara 2"}]}	task	[ "QUESTION_ANSWERING" ]	44,589
facebook/m2m100-12B-avg-10-ckpt	facebook	text2text-generation	[ "transformers", "pytorch", "m2m_100", "text2text-generation", "m2m100-12B", "multilingual", "af", "am", "ar", "ast", "az", "ba", "be", "bg", "bn", "br", "bs", "ca", "ceb", "cs", "cy", "da", "de", "el", "en", "es", "et", "fa", "ff", "fi", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "he", "hi", "hr", "ht", "hu", "hy", "id", "ig", "ilo", "is", "it", "ja", "jv", "ka", "kk", "km", "kn", "ko", "lb", "lg", "ln", "lo", "lt", "lv", "mg", "mk", "ml", "mn", "mr", "ms", "my", "ne", "nl", "no", "ns", "oc", "or", "pa", "pl", "ps", "pt", "ro", "ru", "sd", "si", "sk", "sl", "so", "sq", "sr", "ss", "su", "sv", "sw", "ta", "th", "tl", "tn", "tr", "uk", "ur", "uz", "vi", "wo", "xh", "yi", "yo", "zh", "zu", "arxiv:2010.11125", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-13T21:10:48Z	2023-01-24T17:03:13+00:00	108	0	--- language: - multilingual - af - am - ar - ast - az - ba - be - bg - bn - br - bs - ca - ceb - cs - cy - da - de - el - en - es - et - fa - ff - fi - fr - fy - ga - gd - gl - gu - ha - he - hi - hr - ht - hu - hy - id - ig - ilo - is - it - ja - jv - ka - kk - km - kn - ko - lb - lg - ln - lo - lt - lv - mg - mk - ml - mn - mr - ms - my - ne - nl - false - ns - oc - or - pa - pl - ps - pt - ro - ru - sd - si - sk - sl - so - sq - sr - ss - su - sv - sw - ta - th - tl - tn - tr - uk - ur - uz - vi - wo - xh - yi - yo - zh - zu license: mit tags: - m2m100-12B --- # M2M100 12B (average of last 10 checkpoints) M2M100 is a multilingual encoder-decoder (seq-to-seq) model trained for Many-to-Many multilingual translation. It was introduced in this [paper](https://arxiv.org/abs/2010.11125) and first released in [this](https://github.com/pytorch/fairseq/tree/master/examples/m2m_100) repository. The model that can directly translate between the 9,900 directions of 100 languages. To translate into a target language, the target language id is forced as the first generated token. To force the target language id as the first generated token, pass the `forced_bos_token_id` parameter to the `generate` method. Note: `M2M100Tokenizer` depends on `sentencepiece`, so make sure to install it before running the example. To install `sentencepiece` run `pip install sentencepiece` ```python from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer hi_text = "जीवन एक चॉकलेट बॉक्स की तरह है।" chinese_text = "生活就像一盒巧克力。" model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") # translate Hindi to French tokenizer.src_lang = "hi" encoded_hi = tokenizer(hi_text, return_tensors="pt") generated_tokens = model.generate(encoded_hi, forced_bos_token_id=tokenizer.get_lang_id("fr")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "La vie est comme une boîte de chocolat." # translate Chinese to English tokenizer.src_lang = "zh" encoded_zh = tokenizer(chinese_text, return_tensors="pt") generated_tokens = model.generate(encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "Life is like a box of chocolate." ``` See the [model hub](https://huggingface.co/models?filter=m2m_100) to look for more fine-tuned versions. ## Languages covered Afrikaans (af), Amharic (am), Arabic (ar), Asturian (ast), Azerbaijani (az), Bashkir (ba), Belarusian (be), Bulgarian (bg), Bengali (bn), Breton (br), Bosnian (bs), Catalan; Valencian (ca), Cebuano (ceb), Czech (cs), Welsh (cy), Danish (da), German (de), Greeek (el), English (en), Spanish (es), Estonian (et), Persian (fa), Fulah (ff), Finnish (fi), French (fr), Western Frisian (fy), Irish (ga), Gaelic; Scottish Gaelic (gd), Galician (gl), Gujarati (gu), Hausa (ha), Hebrew (he), Hindi (hi), Croatian (hr), Haitian; Haitian Creole (ht), Hungarian (hu), Armenian (hy), Indonesian (id), Igbo (ig), Iloko (ilo), Icelandic (is), Italian (it), Japanese (ja), Javanese (jv), Georgian (ka), Kazakh (kk), Central Khmer (km), Kannada (kn), Korean (ko), Luxembourgish; Letzeburgesch (lb), Ganda (lg), Lingala (ln), Lao (lo), Lithuanian (lt), Latvian (lv), Malagasy (mg), Macedonian (mk), Malayalam (ml), Mongolian (mn), Marathi (mr), Malay (ms), Burmese (my), Nepali (ne), Dutch; Flemish (nl), Norwegian (no), Northern Sotho (ns), Occitan (post 1500) (oc), Oriya (or), Panjabi; Punjabi (pa), Polish (pl), Pushto; Pashto (ps), Portuguese (pt), Romanian; Moldavian; Moldovan (ro), Russian (ru), Sindhi (sd), Sinhala; Sinhalese (si), Slovak (sk), Slovenian (sl), Somali (so), Albanian (sq), Serbian (sr), Swati (ss), Sundanese (su), Swedish (sv), Swahili (sw), Tamil (ta), Thai (th), Tagalog (tl), Tswana (tn), Turkish (tr), Ukrainian (uk), Urdu (ur), Uzbek (uz), Vietnamese (vi), Wolof (wo), Xhosa (xh), Yiddish (yi), Yoruba (yo), Chinese (zh), Zulu (zu) ## BibTeX entry and citation info ``` @misc{fan2020englishcentric, title={Beyond English-Centric Multilingual Machine Translation}, author={Angela Fan and Shruti Bhosale and Holger Schwenk and Zhiyi Ma and Ahmed El-Kishky and Siddharth Goyal and Mandeep Baines and Onur Celebi and Guillaume Wenzek and Vishrav Chaudhary and Naman Goyal and Tom Birch and Vitaliy Liptchinsky and Sergey Edunov and Edouard Grave and Michael Auli and Armand Joulin}, year={2020}, eprint={2010.11125}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```	null	Non_BioNLP	# M2M100 12B (average of last 10 checkpoints) M2M100 is a multilingual encoder-decoder (seq-to-seq) model trained for Many-to-Many multilingual translation. It was introduced in this [paper](https://arxiv.org/abs/2010.11125) and first released in [this](https://github.com/pytorch/fairseq/tree/master/examples/m2m_100) repository. The model that can directly translate between the 9,900 directions of 100 languages. To translate into a target language, the target language id is forced as the first generated token. To force the target language id as the first generated token, pass the `forced_bos_token_id` parameter to the `generate` method. Note: `M2M100Tokenizer` depends on `sentencepiece`, so make sure to install it before running the example. To install `sentencepiece` run `pip install sentencepiece` ```python from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer hi_text = "जीवन एक चॉकलेट बॉक्स की तरह है।" chinese_text = "生活就像一盒巧克力。" model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") # translate Hindi to French tokenizer.src_lang = "hi" encoded_hi = tokenizer(hi_text, return_tensors="pt") generated_tokens = model.generate(encoded_hi, forced_bos_token_id=tokenizer.get_lang_id("fr")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "La vie est comme une boîte de chocolat." # translate Chinese to English tokenizer.src_lang = "zh" encoded_zh = tokenizer(chinese_text, return_tensors="pt") generated_tokens = model.generate(encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "Life is like a box of chocolate." ``` See the [model hub](https://huggingface.co/models?filter=m2m_100) to look for more fine-tuned versions. ## Languages covered Afrikaans (af), Amharic (am), Arabic (ar), Asturian (ast), Azerbaijani (az), Bashkir (ba), Belarusian (be), Bulgarian (bg), Bengali (bn), Breton (br), Bosnian (bs), Catalan; Valencian (ca), Cebuano (ceb), Czech (cs), Welsh (cy), Danish (da), German (de), Greeek (el), English (en), Spanish (es), Estonian (et), Persian (fa), Fulah (ff), Finnish (fi), French (fr), Western Frisian (fy), Irish (ga), Gaelic; Scottish Gaelic (gd), Galician (gl), Gujarati (gu), Hausa (ha), Hebrew (he), Hindi (hi), Croatian (hr), Haitian; Haitian Creole (ht), Hungarian (hu), Armenian (hy), Indonesian (id), Igbo (ig), Iloko (ilo), Icelandic (is), Italian (it), Japanese (ja), Javanese (jv), Georgian (ka), Kazakh (kk), Central Khmer (km), Kannada (kn), Korean (ko), Luxembourgish; Letzeburgesch (lb), Ganda (lg), Lingala (ln), Lao (lo), Lithuanian (lt), Latvian (lv), Malagasy (mg), Macedonian (mk), Malayalam (ml), Mongolian (mn), Marathi (mr), Malay (ms), Burmese (my), Nepali (ne), Dutch; Flemish (nl), Norwegian (no), Northern Sotho (ns), Occitan (post 1500) (oc), Oriya (or), Panjabi; Punjabi (pa), Polish (pl), Pushto; Pashto (ps), Portuguese (pt), Romanian; Moldavian; Moldovan (ro), Russian (ru), Sindhi (sd), Sinhala; Sinhalese (si), Slovak (sk), Slovenian (sl), Somali (so), Albanian (sq), Serbian (sr), Swati (ss), Sundanese (su), Swedish (sv), Swahili (sw), Tamil (ta), Thai (th), Tagalog (tl), Tswana (tn), Turkish (tr), Ukrainian (uk), Urdu (ur), Uzbek (uz), Vietnamese (vi), Wolof (wo), Xhosa (xh), Yiddish (yi), Yoruba (yo), Chinese (zh), Zulu (zu) ## BibTeX entry and citation info ``` @misc{fan2020englishcentric, title={Beyond English-Centric Multilingual Machine Translation}, author={Angela Fan and Shruti Bhosale and Holger Schwenk and Zhiyi Ma and Ahmed El-Kishky and Siddharth Goyal and Mandeep Baines and Onur Celebi and Guillaume Wenzek and Vishrav Chaudhary and Naman Goyal and Tom Birch and Vitaliy Liptchinsky and Sergey Edunov and Edouard Grave and Michael Auli and Armand Joulin}, year={2020}, eprint={2010.11125}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```	{"language": ["multilingual", "af", "am", "ar", "ast", "az", "ba", "be", "bg", "bn", "br", "bs", "ca", "ceb", "cs", "cy", "da", "de", "el", "en", "es", "et", "fa", "ff", "fi", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "he", "hi", "hr", "ht", "hu", "hy", "id", "ig", "ilo", "is", "it", "ja", "jv", "ka", "kk", "km", "kn", "ko", "lb", "lg", "ln", "lo", "lt", "lv", "mg", "mk", "ml", "mn", "mr", "ms", "my", "ne", "nl", false, "ns", "oc", "or", "pa", "pl", "ps", "pt", "ro", "ru", "sd", "si", "sk", "sl", "so", "sq", "sr", "ss", "su", "sv", "sw", "ta", "th", "tl", "tn", "tr", "uk", "ur", "uz", "vi", "wo", "xh", "yi", "yo", "zh", "zu"], "license": "mit", "tags": ["m2m100-12B"]}	task	[ "TRANSLATION" ]	44,590
Xfgll/RuleGPT-en-grammar	Xfgll	text-generation	[ "transformers", "safetensors", "qwen", "text-generation", "custom_code", "zh", "en", "arxiv:2309.16609", "arxiv:2305.08322", "arxiv:2009.03300", "arxiv:2305.05280", "arxiv:2210.03629", "autotrain_compatible", "region:us" ]	2024-07-15T09:47:31Z	2024-07-15T11:01:49+00:00	15	0	--- language: - zh - en pipeline_tag: text-generation tags: - qwen inference: false --- # Qwen-7B-Chat <p align="center"> <img src="https://qianwen-res.oss-cn-beijing.aliyuncs.com/logo_qwen.jpg" width="400"/> <p> <br> <p align="center"> 🤗 <a href="https://huggingface.co/Qwen">Hugging Face</a>&nbsp&nbsp \| &nbsp&nbsp🤖 <a href="https://modelscope.cn/organization/qwen">ModelScope</a>&nbsp&nbsp \| &nbsp&nbsp 📑 <a href="https://arxiv.org/abs/2309.16609">Paper</a> &nbsp&nbsp ｜ &nbsp&nbsp🖥️ <a href="https://modelscope.cn/studios/qwen/Qwen-7B-Chat-Demo/summary">Demo</a> <br> <a href="assets/wechat.png">WeChat (微信)</a>&nbsp&nbsp \| &nbsp&nbsp<a href="https://discord.gg/z3GAxXZ9Ce">Discord</a>&nbsp&nbsp ｜ &nbsp&nbsp<a href="https://dashscope.aliyun.com">API</a> </p> <br> ## 介绍（Introduction）通义千问-7B（Qwen-7B）是阿里云研发的通义千问大模型系列的70亿参数规模的模型。Qwen-7B是基于Transformer的大语言模型, 在超大规模的预训练数据上进行训练得到。预训练数据类型多样，覆盖广泛，包括大量网络文本、专业书籍、代码等。同时，在Qwen-7B的基础上，我们使用对齐机制打造了基于大语言模型的AI助手Qwen-7B-Chat。相较于最初开源的Qwen-7B模型，我们现已将预训练模型和Chat模型更新到效果更优的版本。本仓库为Qwen-7B-Chat的仓库。如果您想了解更多关于通义千问-7B开源模型的细节，我们建议您参阅[GitHub代码库](https://github.com/QwenLM/Qwen)。 Qwen-7B is the 7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Alibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. Additionally, based on the pretrained Qwen-7B, we release Qwen-7B-Chat, a large-model-based AI assistant, which is trained with alignment techniques. Now we have updated both our pretrained and chat models with better performances. This repository is the one for Qwen-7B-Chat. For more details about Qwen, please refer to the [GitHub](https://github.com/QwenLM/Qwen) code repository. <br> ## 要求（Requirements） * python 3.8及以上版本 * pytorch 1.12及以上版本，推荐2.0及以上版本 * 建议使用CUDA 11.4及以上（GPU用户、flash-attention用户等需考虑此选项） * python 3.8 and above * pytorch 1.12 and above, 2.0 and above are recommended * CUDA 11.4 and above are recommended (this is for GPU users, flash-attention users, etc.) <br> ## 依赖项（Dependency）运行Qwen-7B-Chat，请确保满足上述要求，再执行以下pip命令安装依赖库 To run Qwen-7B-Chat, please make sure you meet the above requirements, and then execute the following pip commands to install the dependent libraries. ```bash pip install transformers==4.32.0 accelerate tiktoken einops scipy transformers_stream_generator==0.0.4 peft deepspeed ``` 另外，推荐安装`flash-attention`库（当前已支持flash attention 2），以实现更高的效率和更低的显存占用。 In addition, it is recommended to install the `flash-attention` library (we support flash attention 2 now.) for higher efficiency and lower memory usage. ```bash git clone https://github.com/Dao-AILab/flash-attention cd flash-attention && pip install . # 下方安装可选，安装可能比较缓慢。 # pip install csrc/layer_norm # pip install csrc/rotary ``` <br> ## 快速使用（Quickstart）下面我们展示了一个使用Qwen-7B-Chat模型，进行多轮对话交互的样例： We show an example of multi-turn interaction with Qwen-7B-Chat in the following code: ```python from modelscope import AutoModelForCausalLM, AutoTokenizer from modelscope import GenerationConfig # Note: The default behavior now has injection attack prevention off. tokenizer = AutoTokenizer.from_pretrained("qwen/Qwen-7B-Chat", trust_remote_code=True) # use bf16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, bf16=True).eval() # use fp16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, fp16=True).eval() # use cpu only # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="cpu", trust_remote_code=True).eval() # use auto mode, automatically select precision based on the device. model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True).eval() # Specify hyperparameters for generation. But if you use transformers>=4.32.0, there is no need to do this. # model.generation_config = GenerationConfig.from_pretrained("Qwen/Qwen-7B-Chat", trust_remote_code=True) # 可指定不同的生成长度、top_p等相关超参 # 第一轮对话 1st dialogue turn response, history = model.chat(tokenizer, "你好", history=None) print(response) # 你好！很高兴为你提供帮助。 # 第二轮对话 2nd dialogue turn response, history = model.chat(tokenizer, "给我讲一个年轻人奋斗创业最终取得成功的故事。", history=history) print(response) # 这是一个关于一个年轻人奋斗创业最终取得成功的故事。 # 故事的主人公叫李明，他来自一个普通的家庭，父母都是普通的工人。从小，李明就立下了一个目标：要成为一名成功的企业家。 # 为了实现这个目标，李明勤奋学习，考上了大学。在大学期间，他积极参加各种创业比赛，获得了不少奖项。他还利用课余时间去实习，积累了宝贵的经验。 # 毕业后，李明决定开始自己的创业之路。他开始寻找投资机会，但多次都被拒绝了。然而，他并没有放弃。他继续努力，不断改进自己的创业计划，并寻找新的投资机会。 # 最终，李明成功地获得了一笔投资，开始了自己的创业之路。他成立了一家科技公司，专注于开发新型软件。在他的领导下，公司迅速发展起来，成为了一家成功的科技企业。 # 李明的成功并不是偶然的。他勤奋、坚韧、勇于冒险，不断学习和改进自己。他的成功也证明了，只要努力奋斗，任何人都有可能取得成功。 # 第三轮对话 3rd dialogue turn response, history = model.chat(tokenizer, "给这个故事起一个标题", history=history) print(response) # 《奋斗创业：一个年轻人的成功之路》 ``` 关于更多的使用说明，请参考我们的[GitHub repo](https://github.com/QwenLM/Qwen)获取更多信息。 For more information, please refer to our [GitHub repo](https://github.com/QwenLM/Qwen) for more information. <br> ## Tokenizer > 注：作为术语的“tokenization”在中文中尚无共识的概念对应，本文档采用英文表达以利说明。基于tiktoken的分词器有别于其他分词器，比如sentencepiece分词器。尤其在微调阶段，需要特别注意特殊token的使用。关于tokenizer的更多信息，以及微调时涉及的相关使用，请参阅[文档](https://github.com/QwenLM/Qwen/blob/main/tokenization_note_zh.md)。 Our tokenizer based on tiktoken is different from other tokenizers, e.g., sentencepiece tokenizer. You need to pay attention to special tokens, especially in finetuning. For more detailed information on the tokenizer and related use in fine-tuning, please refer to the [documentation](https://github.com/QwenLM/Qwen/blob/main/tokenization_note.md). <br> ## 量化 (Quantization) ### 用法 (Usage) 请注意：我们更新量化方案为基于[AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ)的量化，提供Qwen-7B-Chat的Int4量化模型[点击这里](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4)。相比此前方案，该方案在模型评测效果几乎无损，且存储需求更低，推理速度更优。 Note: we provide a new solution based on [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ), and release an Int4 quantized model for Qwen-7B-Chat [Click here](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4), which achieves nearly lossless model effects but improved performance on both memory costs and inference speed, in comparison with the previous solution. 以下我们提供示例说明如何使用Int4量化模型。在开始使用前，请先保证满足要求（如torch 2.0及以上，transformers版本为4.32.0及以上，等等），并安装所需安装包： Here we demonstrate how to use our provided quantized models for inference. Before you start, make sure you meet the requirements of auto-gptq (e.g., torch 2.0 and above, transformers 4.32.0 and above, etc.) and install the required packages: ```bash pip install auto-gptq optimum ``` 如安装`auto-gptq`遇到问题，我们建议您到官方[repo](https://github.com/PanQiWei/AutoGPTQ)搜索合适的预编译wheel。随后即可使用和上述一致的用法调用量化模型： If you meet problems installing `auto-gptq`, we advise you to check out the official [repo](https://github.com/PanQiWei/AutoGPTQ) to find a pre-build wheel. Then you can load the quantized model easily and run inference as same as usual: ```python model = AutoModelForCausalLM.from_pretrained( "Qwen/Qwen-7B-Chat-Int4", device_map="auto", trust_remote_code=True ).eval() response, history = model.chat(tokenizer, "你好", history=None) ``` ### 效果评测我们对BF16，Int8和Int4模型在基准评测上做了测试（使用zero-shot设置），发现量化模型效果损失较小，结果如下所示： We illustrate the zero-shot performance of both BF16, Int8 and Int4 models on the benchmark, and we find that the quantized model does not suffer from significant performance degradation. Results are shown below: \| Quantization \| MMLU \| CEval (val) \| GSM8K \| Humaneval \| \| ------------- \| :--------: \| :----------: \| :----: \| :--------: \| \| BF16 \| 55.8 \| 59.7 \| 50.3 \| 37.2 \| \| Int8 \| 55.4 \| 59.4 \| 48.3 \| 34.8 \| \| Int4 \| 55.1 \| 59.2 \| 49.7 \| 29.9 \| ### 推理速度 (Inference Speed) 我们测算了不同精度模型以及不同FlashAttn库版本下模型生成2048和8192个token的平均推理速度。如图所示： We measured the average inference speed of generating 2048 and 8192 tokens with different quantization levels and versions of flash-attention, respectively. \| Quantization \| FlashAttn \| Speed (2048 tokens) \| Speed (8192 tokens) \| \| ------------- \| :-------: \| :------------------:\| :------------------:\| \| BF16 \| v2 \| 40.93 \| 36.14 \| \| Int8 \| v2 \| 37.47 \| 32.54 \| \| Int4 \| v2 \| 50.09 \| 38.61 \| \| BF16 \| v1 \| 40.75 \| 35.34 \| \| Int8 \| v1 \| 37.51 \| 32.39 \| \| Int4 \| v1 \| 45.98 \| 36.47 \| \| BF16 \| Disabled \| 37.55 \| 33.56 \| \| Int8 \| Disabled \| 37.84 \| 32.65 \| \| Int4 \| Disabled \| 48.12 \| 36.70 \| 具体而言，我们记录在长度为1的上下文的条件下生成8192个token的性能。评测运行于单张A100-SXM4-80G GPU，使用PyTorch 2.0.1和CUDA 11.8。推理速度是生成8192个token的速度均值。 In detail, the setting of profiling is generating 8192 new tokens with 1 context token. The profiling runs on a single A100-SXM4-80G GPU with PyTorch 2.0.1 and CUDA 11.8. The inference speed is averaged over the generated 8192 tokens. 注意：以上Int4/Int8模型生成速度使用autogptq库给出，当前``AutoModelForCausalLM.from_pretrained``载入的模型生成速度会慢大约20%。我们已经将该问题汇报给HuggingFace团队，若有解决方案将即时更新。 Note: The generation speed of the Int4/Int8 models mentioned above is provided by the autogptq library. The current speed of the model loaded using "AutoModelForCausalLM.from_pretrained" will be approximately 20% slower. We have reported this issue to the HuggingFace team and will update it promptly if a solution is available. ### 显存使用 (GPU Memory Usage) 我们还测算了不同模型精度编码2048个token及生成8192个token的峰值显存占用情况。（显存消耗在是否使用FlashAttn的情况下均类似。）结果如下所示： We also profile the peak GPU memory usage for encoding 2048 tokens as context (and generating single token) and generating 8192 tokens (with single token as context) under different quantization levels, respectively. （The GPU memory usage is similar when using flash-attention or not.）The results are shown below. \| Quantization Level \| Peak Usage for Encoding 2048 Tokens \| Peak Usage for Generating 8192 Tokens \| \| ------------------ \| :---------------------------------: \| :-----------------------------------: \| \| BF16 \| 16.99GB \| 22.53GB \| \| Int8 \| 11.20GB \| 16.62GB \| \| Int4 \| 8.21GB \| 13.63GB \| 上述性能测算使用[此脚本](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py)完成。 The above speed and memory profiling are conducted using [this script](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py). <br> ## 模型细节（Model）与Qwen-7B预训练模型相同，Qwen-7B-Chat模型规模基本情况如下所示: The details of the model architecture of Qwen-7B-Chat are listed as follows: \| Hyperparameter \| Value \| \|:----------------\|:------:\| \| n_layers \| 32 \| \| n_heads \| 32 \| \| d_model \| 4096 \| \| vocab size \| 151851 \| \| sequence length \| 8192 \| 在位置编码、FFN激活函数和normalization的实现方式上，我们也采用了目前最流行的做法，即RoPE相对位置编码、SwiGLU激活函数、RMSNorm（可选安装flash-attention加速）。在分词器方面，相比目前主流开源模型以中英词表为主，Qwen-7B-Chat使用了约15万token大小的词表。该词表在GPT-4使用的BPE词表`cl100k_base`基础上，对中文、多语言进行了优化，在对中、英、代码数据的高效编解码的基础上，对部分多语言更加友好，方便用户在不扩展词表的情况下对部分语种进行能力增强。词表对数字按单个数字位切分。调用较为高效的[tiktoken分词库](https://github.com/openai/tiktoken)进行分词。 For position encoding, FFN activation function, and normalization calculation methods, we adopt the prevalent practices, i.e., RoPE relative position encoding, SwiGLU for activation function, and RMSNorm for normalization (optional installation of flash-attention for acceleration). For tokenization, compared to the current mainstream open-source models based on Chinese and English vocabularies, Qwen-7B-Chat uses a vocabulary of over 150K tokens. It first considers efficient encoding of Chinese, English, and code data, and is also more friendly to multilingual languages, enabling users to directly enhance the capability of some languages without expanding the vocabulary. It segments numbers by single digit, and calls the [tiktoken](https://github.com/openai/tiktoken) tokenizer library for efficient tokenization. <br> ## 评测效果（Evaluation）对于Qwen-7B-Chat模型，我们同样评测了常规的中文理解（C-Eval）、英文理解（MMLU）、代码（HumanEval）和数学（GSM8K）等权威任务，同时包含了长序列任务的评测结果。由于Qwen-7B-Chat模型经过对齐后，激发了较强的外部系统调用能力，我们还进行了工具使用能力方面的评测。提示：由于硬件和框架造成的舍入误差，复现结果如有波动属于正常现象。 For Qwen-7B-Chat, we also evaluate the model on C-Eval, MMLU, HumanEval, GSM8K, etc., as well as the benchmark evaluation for long-context understanding, and tool usage. Note: Due to rounding errors caused by hardware and framework, differences in reproduced results are possible. ### 中文评测（Chinese Evaluation） #### C-Eval 在[C-Eval](https://arxiv.org/abs/2305.08322)验证集上，我们评价了Qwen-7B-Chat模型的0-shot & 5-shot准确率 We demonstrate the 0-shot & 5-shot accuracy of Qwen-7B-Chat on C-Eval validation set \| Model \| Avg. Acc. \| \|:--------------------------------:\|:---------:\| \| LLaMA2-7B-Chat \| 31.9 \| \| LLaMA2-13B-Chat \| 36.2 \| \| LLaMA2-70B-Chat \| 44.3 \| \| ChatGLM2-6B-Chat \| 52.6 \| \| InternLM-7B-Chat \| 53.6 \| \| Baichuan2-7B-Chat \| 55.6 \| \| Baichuan2-13B-Chat \| 56.7 \| \| Qwen-7B-Chat (original) (0-shot) \| 54.2 \| \| Qwen-7B-Chat (0-shot) \| 59.7 \| \| Qwen-7B-Chat (5-shot) \| 59.3 \| \| Qwen-14B-Chat (0-shot) \| 69.8 \| \| Qwen-14B-Chat (5-shot) \| 71.7 \| C-Eval测试集上，Qwen-7B-Chat模型的zero-shot准确率结果如下： The zero-shot accuracy of Qwen-7B-Chat on C-Eval testing set is provided below: \| Model \| Avg. \| STEM \| Social Sciences \| Humanities \| Others \| \| :---------------------- \| :------: \| :--: \| :-------------: \| :--------: \| :----: \| \| Chinese-Alpaca-Plus-13B \| 41.5 \| 36.6 \| 49.7 \| 43.1 \| 41.2 \| \| Chinese-Alpaca-2-7B \| 40.3 \| - \| - \| - \| - \| \| ChatGLM2-6B-Chat \| 50.1 \| 46.4 \| 60.4 \| 50.6 \| 46.9 \| \| Baichuan-13B-Chat \| 51.5 \| 43.7 \| 64.6 \| 56.2 \| 49.2 \| \| Qwen-7B-Chat (original) \| 54.6 \| 47.8 \| 67.6 \| 59.3 \| 50.6 \| \| Qwen-7B-Chat \| 58.6 \| 53.3 \| 72.1 \| 62.8 \| 52.0 \| \| Qwen-14B-Chat \| 69.1 \| 65.1 \| 80.9 \| 71.2 \| 63.4 \| 在7B规模模型上，经过人类指令对齐的Qwen-7B-Chat模型，准确率在同类相近规模模型中仍然处于前列。 Compared with other pretrained models with comparable model size, the human-aligned Qwen-7B-Chat performs well in C-Eval accuracy. ### 英文评测（English Evaluation） #### MMLU [MMLU](https://arxiv.org/abs/2009.03300)评测集上，Qwen-7B-Chat模型的 0-shot & 5-shot 准确率如下，效果同样在同类对齐模型中同样表现较优。 The 0-shot & 5-shot accuracy of Qwen-7B-Chat on MMLU is provided below. The performance of Qwen-7B-Chat still on the top between other human-aligned models with comparable size. \| Model \| Avg. Acc. \| \|:--------------------------------:\|:---------:\| \| ChatGLM2-6B-Chat \| 46.0 \| \| LLaMA2-7B-Chat \| 46.2 \| \| InternLM-7B-Chat \| 51.1 \| \| Baichuan2-7B-Chat \| 52.9 \| \| LLaMA2-13B-Chat \| 54.6 \| \| Baichuan2-13B-Chat \| 57.3 \| \| LLaMA2-70B-Chat \| 63.8 \| \| Qwen-7B-Chat (original) (0-shot) \| 53.9 \| \| Qwen-7B-Chat (0-shot) \| 55.8 \| \| Qwen-7B-Chat (5-shot) \| 57.0 \| \| Qwen-14B-Chat (0-shot) \| 64.6 \| \| Qwen-14B-Chat (5-shot) \| 66.5 \| ### 代码评测（Coding Evaluation） Qwen-7B-Chat在[HumanEval](https://github.com/openai/human-eval)的zero-shot Pass@1效果如下 The zero-shot Pass@1 of Qwen-7B-Chat on [HumanEval](https://github.com/openai/human-eval) is demonstrated below \| Model \| Pass@1 \| \|:-----------------------:\|:--------:\| \| ChatGLM2-6B-Chat \| 11.0 \| \| LLaMA2-7B-Chat \| 12.2 \| \| Baichuan2-7B-Chat \| 13.4 \| \| InternLM-7B-Chat \| 14.6 \| \| Baichuan2-13B-Chat \| 17.7 \| \| LLaMA2-13B-Chat \| 18.9 \| \| LLaMA2-70B-Chat \| 32.3 \| \| Qwen-7B-Chat (original) \| 24.4 \| \| Qwen-7B-Chat \| 37.2 \| \| Qwen-14B-Chat \| 43.9 \| ### 数学评测（Mathematics Evaluation）在评测数学能力的[GSM8K](https://github.com/openai/grade-school-math)上，Qwen-7B-Chat的准确率结果如下 The accuracy of Qwen-7B-Chat on GSM8K is shown below \| Model \| Acc. \| \|:------------------------------------:\|:--------:\| \| LLaMA2-7B-Chat \| 26.3 \| \| ChatGLM2-6B-Chat \| 28.8 \| \| Baichuan2-7B-Chat \| 32.8 \| \| InternLM-7B-Chat \| 33.0 \| \| LLaMA2-13B-Chat \| 37.1 \| \| Baichuan2-13B-Chat \| 55.3 \| \| LLaMA2-70B-Chat \| 59.3 \| \| Qwen-7B-Chat (original) (0-shot) \| 41.1 \| \| Qwen-7B-Chat (0-shot) \| 50.3 \| \| Qwen-7B-Chat (8-shot) \| 54.1 \| \| Qwen-14B-Chat (0-shot) \| 60.1 \| \| Qwen-14B-Chat (8-shot) \| 59.3 \| ### 长序列评测（Long-Context Understanding）通过NTK插值，LogN注意力缩放可以扩展Qwen-7B-Chat的上下文长度。在长文本摘要数据集[VCSUM](https://arxiv.org/abs/2305.05280)上（文本平均长度在15K左右），Qwen-7B-Chat的Rouge-L结果如下： (若要启用这些技巧，请将config.json里的`use_dynamic_ntk`和`use_logn_attn`设置为true) We introduce NTK-aware interpolation, LogN attention scaling to extend the context length of Qwen-7B-Chat. The Rouge-L results of Qwen-7B-Chat on long-text summarization dataset [VCSUM](https://arxiv.org/abs/2305.05280) (The average length of this dataset is around 15K) are shown below: (To use these tricks, please set `use_dynamic_ntk` and `use_long_attn` to true in config.json.) \| Model \| VCSUM (zh) \| \|:------------------\|:----------:\| \| GPT-3.5-Turbo-16k \| 16.0 \| \| LLama2-7B-Chat \| 0.2 \| \| InternLM-7B-Chat \| 13.0 \| \| ChatGLM2-6B-Chat \| 16.3 \| \| Qwen-7B-Chat \| 16.6 \| ### 工具使用能力的评测（Tool Usage） #### ReAct Prompting 千问支持通过 [ReAct Prompting](https://arxiv.org/abs/2210.03629) 调用插件/工具/API。ReAct 也是 [LangChain](https://python.langchain.com/) 框架采用的主要方式之一。在我们开源的、用于评估工具使用能力的评测基准上，千问的表现如下： Qwen-Chat supports calling plugins/tools/APIs through [ReAct Prompting](https://arxiv.org/abs/2210.03629). ReAct is also one of the main approaches used by the [LangChain](https://python.langchain.com/) framework. In our evaluation benchmark for assessing tool usage capabilities, Qwen-Chat's performance is as follows: <table> <tr> <th colspan="4" align="center">Chinese Tool-Use Benchmark</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection (Acc.↑)</th><th align="center">Tool Input (Rouge-L↑)</th><th align="center">False Positive Error↓</th> </tr> <tr> <td>GPT-4</td><td align="center">95%</td><td align="center">0.90</td><td align="center">15.0%</td> </tr> <tr> <td>GPT-3.5</td><td align="center">85%</td><td align="center">0.88</td><td align="center">75.0%</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">98%</td><td align="center">0.91</td><td align="center">7.3%</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">98%</td><td align="center">0.93</td><td align="center">2.4%</td> </tr> </table> > 评测基准中出现的插件均没有出现在千问的训练集中。该基准评估了模型在多个候选插件中选择正确插件的准确率、传入插件的参数的合理性、以及假阳率。假阳率（False Positive）定义：在处理不该调用插件的请求时，错误地调用了插件。 > The plugins that appear in the evaluation set do not appear in the training set of Qwen. This benchmark evaluates the accuracy of the model in selecting the correct plugin from multiple candidate plugins, the rationality of the parameters passed into the plugin, and the false positive rate. False Positive: Incorrectly invoking a plugin when it should not have been called when responding to a query. ![](assets/react_showcase_001.png) ![](assets/react_showcase_002.png) #### Code Interpreter 为了考察Qwen使用Python Code Interpreter完成数学解题、数据可视化、及文件处理与爬虫等任务的能力，我们专门建设并开源了一个评测这方面能力的[评测基准](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark)。我们发现Qwen在生成代码的可执行率、结果正确性上均表现较好： To assess Qwen's ability to use the Python Code Interpreter for tasks such as mathematical problem solving, data visualization, and other general-purpose tasks such as file handling and web scraping, we have created and open-sourced a benchmark specifically designed for evaluating these capabilities. You can find the benchmark at this [link](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark). We have observed that Qwen performs well in terms of code executability and result accuracy when generating code: <table> <tr> <th colspan="4" align="center">Executable Rate of Generated Code (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization↑</th><th align="center">General↑</th> </tr> <tr> <td>GPT-4</td><td align="center">91.9</td><td align="center">85.9</td><td align="center">82.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">89.2</td><td align="center">65.0</td><td align="center">74.1</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">41.9</td> <td align="center">33.1</td> <td align="center">24.1 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">50.0</td> <td align="center">40.5</td> <td align="center">48.3 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">85.1</td> <td align="center">54.0</td> <td align="center">70.7 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">93.2</td> <td align="center">55.8</td> <td align="center">74.1 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">78.4</td> <td align="center">44.2</td> <td align="center">62.1 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">70.3</td> <td align="center">44.2</td> <td align="center">65.5 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">82.4</td> <td align="center">64.4</td> <td align="center">67.2 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">89.2</td> <td align="center">84.1</td> <td align="center">65.5</td> </tr> </table> <table> <tr> <th colspan="4" align="center">Accuracy of Code Execution Results (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization-Hard↑</th><th align="center">Visualization-Easy↑</th> </tr> <tr> <td>GPT-4</td><td align="center">82.8</td><td align="center">66.7</td><td align="center">60.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">47.3</td><td align="center">33.3</td><td align="center">55.7</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">3.9</td> <td align="center">14.3</td> <td align="center">39.2 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">8.3</td> <td align="center">8.3</td> <td align="center">40.5 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">14.3</td> <td align="center">26.2</td> <td align="center">60.8 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">28.2</td> <td align="center">27.4</td> <td align="center">62.0 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">28.5</td> <td align="center">4.8</td> <td align="center">40.5 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">34.6</td> <td align="center">21.4</td> <td align="center">45.6 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">41.9</td> <td align="center">40.5</td> <td align="center">54.4 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">58.4</td> <td align="center">53.6</td> <td align="center">59.5</td> </tr> </table> <p align="center"> <br> <img src="assets/code_interpreter_showcase_001.jpg" /> <br> <p> #### Huggingface Agent 千问还具备作为 [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents) 的能力。它在 Huggingface 提供的run模式评测基准上的表现如下： Qwen-Chat also has the capability to be used as a [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents). Its performance on the run-mode benchmark provided by HuggingFace is as follows: <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark- Run Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">100</td><td align="center">100</td><td align="center">97.4</td> </tr> <tr> <td>GPT-3.5</td><td align="center">95.4</td><td align="center">96.3</td><td align="center">87.0</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">86.1</td><td align="center">87.0</td><td align="center">68.9</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">87.0</td><td align="center">88.0</td><td align="center">68.9</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">87.0</td><td align="center">87.0</td><td align="center">71.5</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">93.5</td><td align="center">94.4</td><td align="center">87.0</td> </tr> </table> <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark - Chat Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">98.5</td> </tr> <tr> <td>GPT-3.5</td><td align="center">97.3</td><td align="center">96.8</td><td align="center">89.6</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">91.1</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">89.6</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">94.7</td><td align="center">94.7</td><td align="center">85.1</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">95.5</td> </tr> </table> <br> ## FAQ 如遇到问题，敬请查阅[FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ_zh.md)以及issue区，如仍无法解决再提交issue。 If you meet problems, please refer to [FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ.md) and the issues first to search a solution before you launch a new issue. <br> ## 引用 (Citation) 如果你觉得我们的工作对你有帮助，欢迎引用！ If you find our work helpful, feel free to give us a cite. ``` @article{qwen, title={Qwen Technical Report}, author={Jinze Bai and Shuai Bai and Yunfei Chu and Zeyu Cui and Kai Dang and Xiaodong Deng and Yang Fan and Wenbin Ge and Yu Han and Fei Huang and Binyuan Hui and Luo Ji and Mei Li and Junyang Lin and Runji Lin and Dayiheng Liu and Gao Liu and Chengqiang Lu and Keming Lu and Jianxin Ma and Rui Men and Xingzhang Ren and Xuancheng Ren and Chuanqi Tan and Sinan Tan and Jianhong Tu and Peng Wang and Shijie Wang and Wei Wang and Shengguang Wu and Benfeng Xu and Jin Xu and An Yang and Hao Yang and Jian Yang and Shusheng Yang and Yang Yao and Bowen Yu and Hongyi Yuan and Zheng Yuan and Jianwei Zhang and Xingxuan Zhang and Yichang Zhang and Zhenru Zhang and Chang Zhou and Jingren Zhou and Xiaohuan Zhou and Tianhang Zhu}, journal={arXiv preprint arXiv:2309.16609}, year={2023} } ``` <br> ## 使用协议（License Agreement）我们的代码和模型权重对学术研究完全开放，并支持商用。请查看[LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT)了解具体的开源协议细节。如需商用，请填写[问卷](https://dashscope.console.aliyun.com/openModelApply/qianwen)申请。 Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check [LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT) for more details about the license. If you have requirements for commercial use, please fill out the [form](https://dashscope.console.aliyun.com/openModelApply/qianwen) to apply. <br> ## 联系我们（Contact Us）如果你想给我们的研发团队和产品团队留言，欢迎加入我们的微信群、钉钉群以及Discord！同时，也欢迎通过邮件（[email protected]）联系我们。 If you are interested to leave a message to either our research team or product team, join our Discord or WeChat groups! Also, feel free to send an email to [email protected].	null	Non_BioNLP	# Qwen-7B-Chat <p align="center"> <img src="https://qianwen-res.oss-cn-beijing.aliyuncs.com/logo_qwen.jpg" width="400"/> <p> <br> <p align="center"> 🤗 <a href="https://huggingface.co/Qwen">Hugging Face</a>&nbsp&nbsp \| &nbsp&nbsp🤖 <a href="https://modelscope.cn/organization/qwen">ModelScope</a>&nbsp&nbsp \| &nbsp&nbsp 📑 <a href="https://arxiv.org/abs/2309.16609">Paper</a> &nbsp&nbsp ｜ &nbsp&nbsp🖥️ <a href="https://modelscope.cn/studios/qwen/Qwen-7B-Chat-Demo/summary">Demo</a> <br> <a href="assets/wechat.png">WeChat (微信)</a>&nbsp&nbsp \| &nbsp&nbsp<a href="https://discord.gg/z3GAxXZ9Ce">Discord</a>&nbsp&nbsp ｜ &nbsp&nbsp<a href="https://dashscope.aliyun.com">API</a> </p> <br> ## 介绍（Introduction）通义千问-7B（Qwen-7B）是阿里云研发的通义千问大模型系列的70亿参数规模的模型。Qwen-7B是基于Transformer的大语言模型, 在超大规模的预训练数据上进行训练得到。预训练数据类型多样，覆盖广泛，包括大量网络文本、专业书籍、代码等。同时，在Qwen-7B的基础上，我们使用对齐机制打造了基于大语言模型的AI助手Qwen-7B-Chat。相较于最初开源的Qwen-7B模型，我们现已将预训练模型和Chat模型更新到效果更优的版本。本仓库为Qwen-7B-Chat的仓库。如果您想了解更多关于通义千问-7B开源模型的细节，我们建议您参阅[GitHub代码库](https://github.com/QwenLM/Qwen)。 Qwen-7B is the 7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Alibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. Additionally, based on the pretrained Qwen-7B, we release Qwen-7B-Chat, a large-model-based AI assistant, which is trained with alignment techniques. Now we have updated both our pretrained and chat models with better performances. This repository is the one for Qwen-7B-Chat. For more details about Qwen, please refer to the [GitHub](https://github.com/QwenLM/Qwen) code repository. <br> ## 要求（Requirements） * python 3.8及以上版本 * pytorch 1.12及以上版本，推荐2.0及以上版本 * 建议使用CUDA 11.4及以上（GPU用户、flash-attention用户等需考虑此选项） * python 3.8 and above * pytorch 1.12 and above, 2.0 and above are recommended * CUDA 11.4 and above are recommended (this is for GPU users, flash-attention users, etc.) <br> ## 依赖项（Dependency）运行Qwen-7B-Chat，请确保满足上述要求，再执行以下pip命令安装依赖库 To run Qwen-7B-Chat, please make sure you meet the above requirements, and then execute the following pip commands to install the dependent libraries. ```bash pip install transformers==4.32.0 accelerate tiktoken einops scipy transformers_stream_generator==0.0.4 peft deepspeed ``` 另外，推荐安装`flash-attention`库（当前已支持flash attention 2），以实现更高的效率和更低的显存占用。 In addition, it is recommended to install the `flash-attention` library (we support flash attention 2 now.) for higher efficiency and lower memory usage. ```bash git clone https://github.com/Dao-AILab/flash-attention cd flash-attention && pip install . # 下方安装可选，安装可能比较缓慢。 # pip install csrc/layer_norm # pip install csrc/rotary ``` <br> ## 快速使用（Quickstart）下面我们展示了一个使用Qwen-7B-Chat模型，进行多轮对话交互的样例： We show an example of multi-turn interaction with Qwen-7B-Chat in the following code: ```python from modelscope import AutoModelForCausalLM, AutoTokenizer from modelscope import GenerationConfig # Note: The default behavior now has injection attack prevention off. tokenizer = AutoTokenizer.from_pretrained("qwen/Qwen-7B-Chat", trust_remote_code=True) # use bf16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, bf16=True).eval() # use fp16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, fp16=True).eval() # use cpu only # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="cpu", trust_remote_code=True).eval() # use auto mode, automatically select precision based on the device. model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True).eval() # Specify hyperparameters for generation. But if you use transformers>=4.32.0, there is no need to do this. # model.generation_config = GenerationConfig.from_pretrained("Qwen/Qwen-7B-Chat", trust_remote_code=True) # 可指定不同的生成长度、top_p等相关超参 # 第一轮对话 1st dialogue turn response, history = model.chat(tokenizer, "你好", history=None) print(response) # 你好！很高兴为你提供帮助。 # 第二轮对话 2nd dialogue turn response, history = model.chat(tokenizer, "给我讲一个年轻人奋斗创业最终取得成功的故事。", history=history) print(response) # 这是一个关于一个年轻人奋斗创业最终取得成功的故事。 # 故事的主人公叫李明，他来自一个普通的家庭，父母都是普通的工人。从小，李明就立下了一个目标：要成为一名成功的企业家。 # 为了实现这个目标，李明勤奋学习，考上了大学。在大学期间，他积极参加各种创业比赛，获得了不少奖项。他还利用课余时间去实习，积累了宝贵的经验。 # 毕业后，李明决定开始自己的创业之路。他开始寻找投资机会，但多次都被拒绝了。然而，他并没有放弃。他继续努力，不断改进自己的创业计划，并寻找新的投资机会。 # 最终，李明成功地获得了一笔投资，开始了自己的创业之路。他成立了一家科技公司，专注于开发新型软件。在他的领导下，公司迅速发展起来，成为了一家成功的科技企业。 # 李明的成功并不是偶然的。他勤奋、坚韧、勇于冒险，不断学习和改进自己。他的成功也证明了，只要努力奋斗，任何人都有可能取得成功。 # 第三轮对话 3rd dialogue turn response, history = model.chat(tokenizer, "给这个故事起一个标题", history=history) print(response) # 《奋斗创业：一个年轻人的成功之路》 ``` 关于更多的使用说明，请参考我们的[GitHub repo](https://github.com/QwenLM/Qwen)获取更多信息。 For more information, please refer to our [GitHub repo](https://github.com/QwenLM/Qwen) for more information. <br> ## Tokenizer > 注：作为术语的“tokenization”在中文中尚无共识的概念对应，本文档采用英文表达以利说明。基于tiktoken的分词器有别于其他分词器，比如sentencepiece分词器。尤其在微调阶段，需要特别注意特殊token的使用。关于tokenizer的更多信息，以及微调时涉及的相关使用，请参阅[文档](https://github.com/QwenLM/Qwen/blob/main/tokenization_note_zh.md)。 Our tokenizer based on tiktoken is different from other tokenizers, e.g., sentencepiece tokenizer. You need to pay attention to special tokens, especially in finetuning. For more detailed information on the tokenizer and related use in fine-tuning, please refer to the [documentation](https://github.com/QwenLM/Qwen/blob/main/tokenization_note.md). <br> ## 量化 (Quantization) ### 用法 (Usage) 请注意：我们更新量化方案为基于[AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ)的量化，提供Qwen-7B-Chat的Int4量化模型[点击这里](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4)。相比此前方案，该方案在模型评测效果几乎无损，且存储需求更低，推理速度更优。 Note: we provide a new solution based on [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ), and release an Int4 quantized model for Qwen-7B-Chat [Click here](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4), which achieves nearly lossless model effects but improved performance on both memory costs and inference speed, in comparison with the previous solution. 以下我们提供示例说明如何使用Int4量化模型。在开始使用前，请先保证满足要求（如torch 2.0及以上，transformers版本为4.32.0及以上，等等），并安装所需安装包： Here we demonstrate how to use our provided quantized models for inference. Before you start, make sure you meet the requirements of auto-gptq (e.g., torch 2.0 and above, transformers 4.32.0 and above, etc.) and install the required packages: ```bash pip install auto-gptq optimum ``` 如安装`auto-gptq`遇到问题，我们建议您到官方[repo](https://github.com/PanQiWei/AutoGPTQ)搜索合适的预编译wheel。随后即可使用和上述一致的用法调用量化模型： If you meet problems installing `auto-gptq`, we advise you to check out the official [repo](https://github.com/PanQiWei/AutoGPTQ) to find a pre-build wheel. Then you can load the quantized model easily and run inference as same as usual: ```python model = AutoModelForCausalLM.from_pretrained( "Qwen/Qwen-7B-Chat-Int4", device_map="auto", trust_remote_code=True ).eval() response, history = model.chat(tokenizer, "你好", history=None) ``` ### 效果评测我们对BF16，Int8和Int4模型在基准评测上做了测试（使用zero-shot设置），发现量化模型效果损失较小，结果如下所示： We illustrate the zero-shot performance of both BF16, Int8 and Int4 models on the benchmark, and we find that the quantized model does not suffer from significant performance degradation. Results are shown below: \| Quantization \| MMLU \| CEval (val) \| GSM8K \| Humaneval \| \| ------------- \| :--------: \| :----------: \| :----: \| :--------: \| \| BF16 \| 55.8 \| 59.7 \| 50.3 \| 37.2 \| \| Int8 \| 55.4 \| 59.4 \| 48.3 \| 34.8 \| \| Int4 \| 55.1 \| 59.2 \| 49.7 \| 29.9 \| ### 推理速度 (Inference Speed) 我们测算了不同精度模型以及不同FlashAttn库版本下模型生成2048和8192个token的平均推理速度。如图所示： We measured the average inference speed of generating 2048 and 8192 tokens with different quantization levels and versions of flash-attention, respectively. \| Quantization \| FlashAttn \| Speed (2048 tokens) \| Speed (8192 tokens) \| \| ------------- \| :-------: \| :------------------:\| :------------------:\| \| BF16 \| v2 \| 40.93 \| 36.14 \| \| Int8 \| v2 \| 37.47 \| 32.54 \| \| Int4 \| v2 \| 50.09 \| 38.61 \| \| BF16 \| v1 \| 40.75 \| 35.34 \| \| Int8 \| v1 \| 37.51 \| 32.39 \| \| Int4 \| v1 \| 45.98 \| 36.47 \| \| BF16 \| Disabled \| 37.55 \| 33.56 \| \| Int8 \| Disabled \| 37.84 \| 32.65 \| \| Int4 \| Disabled \| 48.12 \| 36.70 \| 具体而言，我们记录在长度为1的上下文的条件下生成8192个token的性能。评测运行于单张A100-SXM4-80G GPU，使用PyTorch 2.0.1和CUDA 11.8。推理速度是生成8192个token的速度均值。 In detail, the setting of profiling is generating 8192 new tokens with 1 context token. The profiling runs on a single A100-SXM4-80G GPU with PyTorch 2.0.1 and CUDA 11.8. The inference speed is averaged over the generated 8192 tokens. 注意：以上Int4/Int8模型生成速度使用autogptq库给出，当前``AutoModelForCausalLM.from_pretrained``载入的模型生成速度会慢大约20%。我们已经将该问题汇报给HuggingFace团队，若有解决方案将即时更新。 Note: The generation speed of the Int4/Int8 models mentioned above is provided by the autogptq library. The current speed of the model loaded using "AutoModelForCausalLM.from_pretrained" will be approximately 20% slower. We have reported this issue to the HuggingFace team and will update it promptly if a solution is available. ### 显存使用 (GPU Memory Usage) 我们还测算了不同模型精度编码2048个token及生成8192个token的峰值显存占用情况。（显存消耗在是否使用FlashAttn的情况下均类似。）结果如下所示： We also profile the peak GPU memory usage for encoding 2048 tokens as context (and generating single token) and generating 8192 tokens (with single token as context) under different quantization levels, respectively. （The GPU memory usage is similar when using flash-attention or not.）The results are shown below. \| Quantization Level \| Peak Usage for Encoding 2048 Tokens \| Peak Usage for Generating 8192 Tokens \| \| ------------------ \| :---------------------------------: \| :-----------------------------------: \| \| BF16 \| 16.99GB \| 22.53GB \| \| Int8 \| 11.20GB \| 16.62GB \| \| Int4 \| 8.21GB \| 13.63GB \| 上述性能测算使用[此脚本](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py)完成。 The above speed and memory profiling are conducted using [this script](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py). <br> ## 模型细节（Model）与Qwen-7B预训练模型相同，Qwen-7B-Chat模型规模基本情况如下所示: The details of the model architecture of Qwen-7B-Chat are listed as follows: \| Hyperparameter \| Value \| \|:----------------\|:------:\| \| n_layers \| 32 \| \| n_heads \| 32 \| \| d_model \| 4096 \| \| vocab size \| 151851 \| \| sequence length \| 8192 \| 在位置编码、FFN激活函数和normalization的实现方式上，我们也采用了目前最流行的做法，即RoPE相对位置编码、SwiGLU激活函数、RMSNorm（可选安装flash-attention加速）。在分词器方面，相比目前主流开源模型以中英词表为主，Qwen-7B-Chat使用了约15万token大小的词表。该词表在GPT-4使用的BPE词表`cl100k_base`基础上，对中文、多语言进行了优化，在对中、英、代码数据的高效编解码的基础上，对部分多语言更加友好，方便用户在不扩展词表的情况下对部分语种进行能力增强。词表对数字按单个数字位切分。调用较为高效的[tiktoken分词库](https://github.com/openai/tiktoken)进行分词。 For position encoding, FFN activation function, and normalization calculation methods, we adopt the prevalent practices, i.e., RoPE relative position encoding, SwiGLU for activation function, and RMSNorm for normalization (optional installation of flash-attention for acceleration). For tokenization, compared to the current mainstream open-source models based on Chinese and English vocabularies, Qwen-7B-Chat uses a vocabulary of over 150K tokens. It first considers efficient encoding of Chinese, English, and code data, and is also more friendly to multilingual languages, enabling users to directly enhance the capability of some languages without expanding the vocabulary. It segments numbers by single digit, and calls the [tiktoken](https://github.com/openai/tiktoken) tokenizer library for efficient tokenization. <br> ## 评测效果（Evaluation）对于Qwen-7B-Chat模型，我们同样评测了常规的中文理解（C-Eval）、英文理解（MMLU）、代码（HumanEval）和数学（GSM8K）等权威任务，同时包含了长序列任务的评测结果。由于Qwen-7B-Chat模型经过对齐后，激发了较强的外部系统调用能力，我们还进行了工具使用能力方面的评测。提示：由于硬件和框架造成的舍入误差，复现结果如有波动属于正常现象。 For Qwen-7B-Chat, we also evaluate the model on C-Eval, MMLU, HumanEval, GSM8K, etc., as well as the benchmark evaluation for long-context understanding, and tool usage. Note: Due to rounding errors caused by hardware and framework, differences in reproduced results are possible. ### 中文评测（Chinese Evaluation） #### C-Eval 在[C-Eval](https://arxiv.org/abs/2305.08322)验证集上，我们评价了Qwen-7B-Chat模型的0-shot & 5-shot准确率 We demonstrate the 0-shot & 5-shot accuracy of Qwen-7B-Chat on C-Eval validation set \| Model \| Avg. Acc. \| \|:--------------------------------:\|:---------:\| \| LLaMA2-7B-Chat \| 31.9 \| \| LLaMA2-13B-Chat \| 36.2 \| \| LLaMA2-70B-Chat \| 44.3 \| \| ChatGLM2-6B-Chat \| 52.6 \| \| InternLM-7B-Chat \| 53.6 \| \| Baichuan2-7B-Chat \| 55.6 \| \| Baichuan2-13B-Chat \| 56.7 \| \| Qwen-7B-Chat (original) (0-shot) \| 54.2 \| \| Qwen-7B-Chat (0-shot) \| 59.7 \| \| Qwen-7B-Chat (5-shot) \| 59.3 \| \| Qwen-14B-Chat (0-shot) \| 69.8 \| \| Qwen-14B-Chat (5-shot) \| 71.7 \| C-Eval测试集上，Qwen-7B-Chat模型的zero-shot准确率结果如下： The zero-shot accuracy of Qwen-7B-Chat on C-Eval testing set is provided below: \| Model \| Avg. \| STEM \| Social Sciences \| Humanities \| Others \| \| :---------------------- \| :------: \| :--: \| :-------------: \| :--------: \| :----: \| \| Chinese-Alpaca-Plus-13B \| 41.5 \| 36.6 \| 49.7 \| 43.1 \| 41.2 \| \| Chinese-Alpaca-2-7B \| 40.3 \| - \| - \| - \| - \| \| ChatGLM2-6B-Chat \| 50.1 \| 46.4 \| 60.4 \| 50.6 \| 46.9 \| \| Baichuan-13B-Chat \| 51.5 \| 43.7 \| 64.6 \| 56.2 \| 49.2 \| \| Qwen-7B-Chat (original) \| 54.6 \| 47.8 \| 67.6 \| 59.3 \| 50.6 \| \| Qwen-7B-Chat \| 58.6 \| 53.3 \| 72.1 \| 62.8 \| 52.0 \| \| Qwen-14B-Chat \| 69.1 \| 65.1 \| 80.9 \| 71.2 \| 63.4 \| 在7B规模模型上，经过人类指令对齐的Qwen-7B-Chat模型，准确率在同类相近规模模型中仍然处于前列。 Compared with other pretrained models with comparable model size, the human-aligned Qwen-7B-Chat performs well in C-Eval accuracy. ### 英文评测（English Evaluation） #### MMLU [MMLU](https://arxiv.org/abs/2009.03300)评测集上，Qwen-7B-Chat模型的 0-shot & 5-shot 准确率如下，效果同样在同类对齐模型中同样表现较优。 The 0-shot & 5-shot accuracy of Qwen-7B-Chat on MMLU is provided below. The performance of Qwen-7B-Chat still on the top between other human-aligned models with comparable size. \| Model \| Avg. Acc. \| \|:--------------------------------:\|:---------:\| \| ChatGLM2-6B-Chat \| 46.0 \| \| LLaMA2-7B-Chat \| 46.2 \| \| InternLM-7B-Chat \| 51.1 \| \| Baichuan2-7B-Chat \| 52.9 \| \| LLaMA2-13B-Chat \| 54.6 \| \| Baichuan2-13B-Chat \| 57.3 \| \| LLaMA2-70B-Chat \| 63.8 \| \| Qwen-7B-Chat (original) (0-shot) \| 53.9 \| \| Qwen-7B-Chat (0-shot) \| 55.8 \| \| Qwen-7B-Chat (5-shot) \| 57.0 \| \| Qwen-14B-Chat (0-shot) \| 64.6 \| \| Qwen-14B-Chat (5-shot) \| 66.5 \| ### 代码评测（Coding Evaluation） Qwen-7B-Chat在[HumanEval](https://github.com/openai/human-eval)的zero-shot Pass@1效果如下 The zero-shot Pass@1 of Qwen-7B-Chat on [HumanEval](https://github.com/openai/human-eval) is demonstrated below \| Model \| Pass@1 \| \|:-----------------------:\|:--------:\| \| ChatGLM2-6B-Chat \| 11.0 \| \| LLaMA2-7B-Chat \| 12.2 \| \| Baichuan2-7B-Chat \| 13.4 \| \| InternLM-7B-Chat \| 14.6 \| \| Baichuan2-13B-Chat \| 17.7 \| \| LLaMA2-13B-Chat \| 18.9 \| \| LLaMA2-70B-Chat \| 32.3 \| \| Qwen-7B-Chat (original) \| 24.4 \| \| Qwen-7B-Chat \| 37.2 \| \| Qwen-14B-Chat \| 43.9 \| ### 数学评测（Mathematics Evaluation）在评测数学能力的[GSM8K](https://github.com/openai/grade-school-math)上，Qwen-7B-Chat的准确率结果如下 The accuracy of Qwen-7B-Chat on GSM8K is shown below \| Model \| Acc. \| \|:------------------------------------:\|:--------:\| \| LLaMA2-7B-Chat \| 26.3 \| \| ChatGLM2-6B-Chat \| 28.8 \| \| Baichuan2-7B-Chat \| 32.8 \| \| InternLM-7B-Chat \| 33.0 \| \| LLaMA2-13B-Chat \| 37.1 \| \| Baichuan2-13B-Chat \| 55.3 \| \| LLaMA2-70B-Chat \| 59.3 \| \| Qwen-7B-Chat (original) (0-shot) \| 41.1 \| \| Qwen-7B-Chat (0-shot) \| 50.3 \| \| Qwen-7B-Chat (8-shot) \| 54.1 \| \| Qwen-14B-Chat (0-shot) \| 60.1 \| \| Qwen-14B-Chat (8-shot) \| 59.3 \| ### 长序列评测（Long-Context Understanding）通过NTK插值，LogN注意力缩放可以扩展Qwen-7B-Chat的上下文长度。在长文本摘要数据集[VCSUM](https://arxiv.org/abs/2305.05280)上（文本平均长度在15K左右），Qwen-7B-Chat的Rouge-L结果如下： (若要启用这些技巧，请将config.json里的`use_dynamic_ntk`和`use_logn_attn`设置为true) We introduce NTK-aware interpolation, LogN attention scaling to extend the context length of Qwen-7B-Chat. The Rouge-L results of Qwen-7B-Chat on long-text summarization dataset [VCSUM](https://arxiv.org/abs/2305.05280) (The average length of this dataset is around 15K) are shown below: (To use these tricks, please set `use_dynamic_ntk` and `use_long_attn` to true in config.json.) \| Model \| VCSUM (zh) \| \|:------------------\|:----------:\| \| GPT-3.5-Turbo-16k \| 16.0 \| \| LLama2-7B-Chat \| 0.2 \| \| InternLM-7B-Chat \| 13.0 \| \| ChatGLM2-6B-Chat \| 16.3 \| \| Qwen-7B-Chat \| 16.6 \| ### 工具使用能力的评测（Tool Usage） #### ReAct Prompting 千问支持通过 [ReAct Prompting](https://arxiv.org/abs/2210.03629) 调用插件/工具/API。ReAct 也是 [LangChain](https://python.langchain.com/) 框架采用的主要方式之一。在我们开源的、用于评估工具使用能力的评测基准上，千问的表现如下： Qwen-Chat supports calling plugins/tools/APIs through [ReAct Prompting](https://arxiv.org/abs/2210.03629). ReAct is also one of the main approaches used by the [LangChain](https://python.langchain.com/) framework. In our evaluation benchmark for assessing tool usage capabilities, Qwen-Chat's performance is as follows: <table> <tr> <th colspan="4" align="center">Chinese Tool-Use Benchmark</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection (Acc.↑)</th><th align="center">Tool Input (Rouge-L↑)</th><th align="center">False Positive Error↓</th> </tr> <tr> <td>GPT-4</td><td align="center">95%</td><td align="center">0.90</td><td align="center">15.0%</td> </tr> <tr> <td>GPT-3.5</td><td align="center">85%</td><td align="center">0.88</td><td align="center">75.0%</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">98%</td><td align="center">0.91</td><td align="center">7.3%</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">98%</td><td align="center">0.93</td><td align="center">2.4%</td> </tr> </table> > 评测基准中出现的插件均没有出现在千问的训练集中。该基准评估了模型在多个候选插件中选择正确插件的准确率、传入插件的参数的合理性、以及假阳率。假阳率（False Positive）定义：在处理不该调用插件的请求时，错误地调用了插件。 > The plugins that appear in the evaluation set do not appear in the training set of Qwen. This benchmark evaluates the accuracy of the model in selecting the correct plugin from multiple candidate plugins, the rationality of the parameters passed into the plugin, and the false positive rate. False Positive: Incorrectly invoking a plugin when it should not have been called when responding to a query. ![](assets/react_showcase_001.png) ![](assets/react_showcase_002.png) #### Code Interpreter 为了考察Qwen使用Python Code Interpreter完成数学解题、数据可视化、及文件处理与爬虫等任务的能力，我们专门建设并开源了一个评测这方面能力的[评测基准](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark)。我们发现Qwen在生成代码的可执行率、结果正确性上均表现较好： To assess Qwen's ability to use the Python Code Interpreter for tasks such as mathematical problem solving, data visualization, and other general-purpose tasks such as file handling and web scraping, we have created and open-sourced a benchmark specifically designed for evaluating these capabilities. You can find the benchmark at this [link](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark). We have observed that Qwen performs well in terms of code executability and result accuracy when generating code: <table> <tr> <th colspan="4" align="center">Executable Rate of Generated Code (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization↑</th><th align="center">General↑</th> </tr> <tr> <td>GPT-4</td><td align="center">91.9</td><td align="center">85.9</td><td align="center">82.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">89.2</td><td align="center">65.0</td><td align="center">74.1</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">41.9</td> <td align="center">33.1</td> <td align="center">24.1 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">50.0</td> <td align="center">40.5</td> <td align="center">48.3 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">85.1</td> <td align="center">54.0</td> <td align="center">70.7 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">93.2</td> <td align="center">55.8</td> <td align="center">74.1 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">78.4</td> <td align="center">44.2</td> <td align="center">62.1 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">70.3</td> <td align="center">44.2</td> <td align="center">65.5 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">82.4</td> <td align="center">64.4</td> <td align="center">67.2 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">89.2</td> <td align="center">84.1</td> <td align="center">65.5</td> </tr> </table> <table> <tr> <th colspan="4" align="center">Accuracy of Code Execution Results (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization-Hard↑</th><th align="center">Visualization-Easy↑</th> </tr> <tr> <td>GPT-4</td><td align="center">82.8</td><td align="center">66.7</td><td align="center">60.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">47.3</td><td align="center">33.3</td><td align="center">55.7</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">3.9</td> <td align="center">14.3</td> <td align="center">39.2 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">8.3</td> <td align="center">8.3</td> <td align="center">40.5 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">14.3</td> <td align="center">26.2</td> <td align="center">60.8 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">28.2</td> <td align="center">27.4</td> <td align="center">62.0 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">28.5</td> <td align="center">4.8</td> <td align="center">40.5 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">34.6</td> <td align="center">21.4</td> <td align="center">45.6 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">41.9</td> <td align="center">40.5</td> <td align="center">54.4 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">58.4</td> <td align="center">53.6</td> <td align="center">59.5</td> </tr> </table> <p align="center"> <br> <img src="assets/code_interpreter_showcase_001.jpg" /> <br> <p> #### Huggingface Agent 千问还具备作为 [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents) 的能力。它在 Huggingface 提供的run模式评测基准上的表现如下： Qwen-Chat also has the capability to be used as a [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents). Its performance on the run-mode benchmark provided by HuggingFace is as follows: <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark- Run Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">100</td><td align="center">100</td><td align="center">97.4</td> </tr> <tr> <td>GPT-3.5</td><td align="center">95.4</td><td align="center">96.3</td><td align="center">87.0</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">86.1</td><td align="center">87.0</td><td align="center">68.9</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">87.0</td><td align="center">88.0</td><td align="center">68.9</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">87.0</td><td align="center">87.0</td><td align="center">71.5</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">93.5</td><td align="center">94.4</td><td align="center">87.0</td> </tr> </table> <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark - Chat Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">98.5</td> </tr> <tr> <td>GPT-3.5</td><td align="center">97.3</td><td align="center">96.8</td><td align="center">89.6</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">91.1</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">89.6</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">94.7</td><td align="center">94.7</td><td align="center">85.1</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">95.5</td> </tr> </table> <br> ## FAQ 如遇到问题，敬请查阅[FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ_zh.md)以及issue区，如仍无法解决再提交issue。 If you meet problems, please refer to [FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ.md) and the issues first to search a solution before you launch a new issue. <br> ## 引用 (Citation) 如果你觉得我们的工作对你有帮助，欢迎引用！ If you find our work helpful, feel free to give us a cite. ``` @article{qwen, title={Qwen Technical Report}, author={Jinze Bai and Shuai Bai and Yunfei Chu and Zeyu Cui and Kai Dang and Xiaodong Deng and Yang Fan and Wenbin Ge and Yu Han and Fei Huang and Binyuan Hui and Luo Ji and Mei Li and Junyang Lin and Runji Lin and Dayiheng Liu and Gao Liu and Chengqiang Lu and Keming Lu and Jianxin Ma and Rui Men and Xingzhang Ren and Xuancheng Ren and Chuanqi Tan and Sinan Tan and Jianhong Tu and Peng Wang and Shijie Wang and Wei Wang and Shengguang Wu and Benfeng Xu and Jin Xu and An Yang and Hao Yang and Jian Yang and Shusheng Yang and Yang Yao and Bowen Yu and Hongyi Yuan and Zheng Yuan and Jianwei Zhang and Xingxuan Zhang and Yichang Zhang and Zhenru Zhang and Chang Zhou and Jingren Zhou and Xiaohuan Zhou and Tianhang Zhu}, journal={arXiv preprint arXiv:2309.16609}, year={2023} } ``` <br> ## 使用协议（License Agreement）我们的代码和模型权重对学术研究完全开放，并支持商用。请查看[LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT)了解具体的开源协议细节。如需商用，请填写[问卷](https://dashscope.console.aliyun.com/openModelApply/qianwen)申请。 Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check [LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT) for more details about the license. If you have requirements for commercial use, please fill out the [form](https://dashscope.console.aliyun.com/openModelApply/qianwen) to apply. <br> ## 联系我们（Contact Us）如果你想给我们的研发团队和产品团队留言，欢迎加入我们的微信群、钉钉群以及Discord！同时，也欢迎通过邮件（[email protected]）联系我们。 If you are interested to leave a message to either our research team or product team, join our Discord or WeChat groups! Also, feel free to send an email to [email protected].	{"language": ["zh", "en"], "pipeline_tag": "text-generation", "tags": ["qwen"], "inference": false}	task	[ "SUMMARIZATION" ]	44,591
mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF	mradermacher	translation	[ "transformers", "gguf", "mergekit", "merge", "translation", "japanese_media", "otaku_media", "visual_novels", "VNs", "en", "ja", "base_model:Casual-Autopsy/Llama-3-VNTL-Yollisa-8B", "base_model:quantized:Casual-Autopsy/Llama-3-VNTL-Yollisa-8B", "endpoints_compatible", "region:us", "imatrix" ]	2025-03-05T21:26:26Z	2025-03-05T23:56:29+00:00	728	1	--- base_model: Casual-Autopsy/Llama-3-VNTL-Yollisa-8B language: - en - ja library_name: transformers tags: - mergekit - merge - translation - japanese_media - otaku_media - visual_novels - VNs quantized_by: mradermacher --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: nicoboss --> weighted/imatrix quants of https://huggingface.co/Casual-Autopsy/Llama-3-VNTL-Yollisa-8B <!-- provided-files --> static quants are available at https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) \| Link \| Type \| Size/GB \| Notes \| \|:-----\|:-----\|--------:\|:------\| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_S.gguf) \| i1-IQ1_S \| 2.1 \| for the desperate \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_M.gguf) \| i1-IQ1_M \| 2.3 \| mostly desperate \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XXS.gguf) \| i1-IQ2_XXS \| 2.5 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XS.gguf) \| i1-IQ2_XS \| 2.7 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_S.gguf) \| i1-IQ2_S \| 2.9 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_M.gguf) \| i1-IQ2_M \| 3.0 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K_S.gguf) \| i1-Q2_K_S \| 3.1 \| very low quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K.gguf) \| i1-Q2_K \| 3.3 \| IQ3_XXS probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XXS.gguf) \| i1-IQ3_XXS \| 3.4 \| lower quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XS.gguf) \| i1-IQ3_XS \| 3.6 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_S.gguf) \| i1-Q3_K_S \| 3.8 \| IQ3_XS probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_S.gguf) \| i1-IQ3_S \| 3.8 \| beats Q3_K* \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_M.gguf) \| i1-IQ3_M \| 3.9 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_M.gguf) \| i1-Q3_K_M \| 4.1 \| IQ3_S probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_L.gguf) \| i1-Q3_K_L \| 4.4 \| IQ3_M probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_XS.gguf) \| i1-IQ4_XS \| 4.5 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_0.gguf) \| i1-Q4_0 \| 4.8 \| fast, low quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_NL.gguf) \| i1-IQ4_NL \| 4.8 \| prefer IQ4_XS \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_S.gguf) \| i1-Q4_K_S \| 4.8 \| optimal size/speed/quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_M.gguf) \| i1-Q4_K_M \| 5.0 \| fast, recommended \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_1.gguf) \| i1-Q4_1 \| 5.2 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_S.gguf) \| i1-Q5_K_S \| 5.7 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_M.gguf) \| i1-Q5_K_M \| 5.8 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q6_K.gguf) \| i1-Q6_K \| 6.7 \| practically like static Q6_K \| Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better): ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png) And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to. <!-- end -->	null	Non_BioNLP	## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: nicoboss --> weighted/imatrix quants of https://huggingface.co/Casual-Autopsy/Llama-3-VNTL-Yollisa-8B <!-- provided-files --> static quants are available at https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) \| Link \| Type \| Size/GB \| Notes \| \|:-----\|:-----\|--------:\|:------\| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_S.gguf) \| i1-IQ1_S \| 2.1 \| for the desperate \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_M.gguf) \| i1-IQ1_M \| 2.3 \| mostly desperate \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XXS.gguf) \| i1-IQ2_XXS \| 2.5 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XS.gguf) \| i1-IQ2_XS \| 2.7 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_S.gguf) \| i1-IQ2_S \| 2.9 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_M.gguf) \| i1-IQ2_M \| 3.0 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K_S.gguf) \| i1-Q2_K_S \| 3.1 \| very low quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K.gguf) \| i1-Q2_K \| 3.3 \| IQ3_XXS probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XXS.gguf) \| i1-IQ3_XXS \| 3.4 \| lower quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XS.gguf) \| i1-IQ3_XS \| 3.6 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_S.gguf) \| i1-Q3_K_S \| 3.8 \| IQ3_XS probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_S.gguf) \| i1-IQ3_S \| 3.8 \| beats Q3_K* \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_M.gguf) \| i1-IQ3_M \| 3.9 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_M.gguf) \| i1-Q3_K_M \| 4.1 \| IQ3_S probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_L.gguf) \| i1-Q3_K_L \| 4.4 \| IQ3_M probably better \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_XS.gguf) \| i1-IQ4_XS \| 4.5 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_0.gguf) \| i1-Q4_0 \| 4.8 \| fast, low quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_NL.gguf) \| i1-IQ4_NL \| 4.8 \| prefer IQ4_XS \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_S.gguf) \| i1-Q4_K_S \| 4.8 \| optimal size/speed/quality \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_M.gguf) \| i1-Q4_K_M \| 5.0 \| fast, recommended \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_1.gguf) \| i1-Q4_1 \| 5.2 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_S.gguf) \| i1-Q5_K_S \| 5.7 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_M.gguf) \| i1-Q5_K_M \| 5.8 \| \| \| [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q6_K.gguf) \| i1-Q6_K \| 6.7 \| practically like static Q6_K \| Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better): ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png) And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to. <!-- end -->	{"base_model": "Casual-Autopsy/Llama-3-VNTL-Yollisa-8B", "language": ["en", "ja"], "library_name": "transformers", "tags": ["mergekit", "merge", "translation", "japanese_media", "otaku_media", "visual_novels", "VNs"], "quantized_by": "mradermacher"}	task	[ "TRANSLATION" ]	44,592
banglagov/banELECTRA-Base	banglagov	null	[ "pytorch", "electra", "region:us" ]	2024-11-05T04:15:26Z	2025-01-24T04:09:52+00:00	41	0	--- {} --- # Model Card: banELECTRA-Base ## Model Details The benElectra model is a Bangla adaptation of ELECTRA (Efficiently Learning an Encoder that Classifies Token Replacements Accurately), a pre-training method for language models introduced by researchers at Google.ELECTRA uses a unique training strategy called contrastive learning, which differs from traditional masked language modeling (MLM) methods like BERT.After pre-training, only the discriminator is fine-tuned on downstream tasks, making ELECTRA a more efficient alternative to BERT, achieving higher performance with fewer parameters. The banELECTRA-Base model is tailored for Bangla text and fine-tuned for tasks like `Named Entity Recognition (NER), Part-of-Speech (POS) tagging,Sentence Similarity,Paraphrase Identification,etc.`The model was trained on two NVIDIA GeForce A40 GPUs. ## Training Data The banELECTRA-Base model was pre-trained on a 32 GB Bangla text dataset. Below are the dataset statistics: - Total Words: ~1.996 billion - Unique Words: ~21.24 million - Total Sentences: ~165.38 million - Total Documents: ~15.62 million ## Model Architecture and Training The benELECTRA model was trained using the official [ELECTRA repository](https://huggingface.co/docs/transformers/en/model_doc/electra) with carefully selected hyperparameters to optimize performance for Bangla text. The model uses a vocabulary size of 50,000 tokens and consists of 12 hidden layers with 768 hidden dimensions and 12 attention heads in the discriminator. The generator is scaled to one-third the size of the discriminator, and training is conducted with a maximum sequence length of 256. The training employed a batch size of 96, a learning rate of 0.0004 with 10,000 warm-up steps, and a total of 1,000,000 training steps. Regularization techniques, such as a dropout rate of 0.1 and a weight decay of 0.01, were applied to improve generalization. ## How to Use ```bash from transformers import ElectraTokenizer, ElectraForSequenceClassification model_name = "banglagov/banELECTRA-Base" tokenizer = ElectraTokenizer.from_pretrained(model_name) model = ElectraForSequenceClassification.from_pretrained(model_name) text = "এর ফলে আগামী বছর বেকারত্বের হার বৃদ্ধি এবং অর্থনৈতিক মন্দার আশঙ্কায় ইউরোপীয় ইউনিয়ন ।" inputs = tokenizer(text, return_tensors="pt") print("Input Tokens ids:", inputs) ``` ## Experimental Results The banELECTRA-Base model demonstrates strong performance on downstream tasks, as shown below: \| Task \| Precision \| Recall \| F1 \| \|-------------------------\|---------------\|------------\|-----------\| \| Named Entity Recognition (NER) \| 0.8842 \| 0.7930 \| 0.8249 \| \| Part-of-Speech (POS) Tagging \| 0.8757 \| 0.8717 \| 0.8706 \| Here we used banELECTRA-Base model with Noisy Label model architecture.	null	Non_BioNLP	# Model Card: banELECTRA-Base ## Model Details The benElectra model is a Bangla adaptation of ELECTRA (Efficiently Learning an Encoder that Classifies Token Replacements Accurately), a pre-training method for language models introduced by researchers at Google.ELECTRA uses a unique training strategy called contrastive learning, which differs from traditional masked language modeling (MLM) methods like BERT.After pre-training, only the discriminator is fine-tuned on downstream tasks, making ELECTRA a more efficient alternative to BERT, achieving higher performance with fewer parameters. The banELECTRA-Base model is tailored for Bangla text and fine-tuned for tasks like `Named Entity Recognition (NER), Part-of-Speech (POS) tagging,Sentence Similarity,Paraphrase Identification,etc.`The model was trained on two NVIDIA GeForce A40 GPUs. ## Training Data The banELECTRA-Base model was pre-trained on a 32 GB Bangla text dataset. Below are the dataset statistics: - Total Words: ~1.996 billion - Unique Words: ~21.24 million - Total Sentences: ~165.38 million - Total Documents: ~15.62 million ## Model Architecture and Training The benELECTRA model was trained using the official [ELECTRA repository](https://huggingface.co/docs/transformers/en/model_doc/electra) with carefully selected hyperparameters to optimize performance for Bangla text. The model uses a vocabulary size of 50,000 tokens and consists of 12 hidden layers with 768 hidden dimensions and 12 attention heads in the discriminator. The generator is scaled to one-third the size of the discriminator, and training is conducted with a maximum sequence length of 256. The training employed a batch size of 96, a learning rate of 0.0004 with 10,000 warm-up steps, and a total of 1,000,000 training steps. Regularization techniques, such as a dropout rate of 0.1 and a weight decay of 0.01, were applied to improve generalization. ## How to Use ```bash from transformers import ElectraTokenizer, ElectraForSequenceClassification model_name = "banglagov/banELECTRA-Base" tokenizer = ElectraTokenizer.from_pretrained(model_name) model = ElectraForSequenceClassification.from_pretrained(model_name) text = "এর ফলে আগামী বছর বেকারত্বের হার বৃদ্ধি এবং অর্থনৈতিক মন্দার আশঙ্কায় ইউরোপীয় ইউনিয়ন ।" inputs = tokenizer(text, return_tensors="pt") print("Input Tokens ids:", inputs) ``` ## Experimental Results The banELECTRA-Base model demonstrates strong performance on downstream tasks, as shown below: \| Task \| Precision \| Recall \| F1 \| \|-------------------------\|---------------\|------------\|-----------\| \| Named Entity Recognition (NER) \| 0.8842 \| 0.7930 \| 0.8249 \| \| Part-of-Speech (POS) Tagging \| 0.8757 \| 0.8717 \| 0.8706 \| Here we used banELECTRA-Base model with Noisy Label model architecture.	{}	task	[ "NAMED_ENTITY_RECOGNITION" ]	44,593
mini1013/master_cate_ac10	mini1013	text-classification	[ "setfit", "safetensors", "roberta", "sentence-transformers", "text-classification", "generated_from_setfit_trainer", "arxiv:2209.11055", "base_model:mini1013/master_domain", "base_model:finetune:mini1013/master_domain", "model-index", "region:us" ]	2024-11-25T10:33:19Z	2024-11-25T10:33:39+00:00	454	0	--- base_model: mini1013/master_domain library_name: setfit metrics: - metric pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 벤시몽 RAIN BOOTS MID - 7color DOLPHIN GREY_40 260 오리상점 - text: 플레이볼 오리진 뮬 (PLAYBALL ORIGIN MULE) NY (Off White) 화이트_230 주식회사 에프앤에프 - text: XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다 - text: 다이어트 슬리퍼 다리 부종 스트레칭 균형 실내화 핑크 33-37_33 글로벌다이렉트 - text: 케즈 챔피온 스트랩 캔버스5 M01778F001 Black/Black/Black_230 블루빌리 inference: true model-index: - name: SetFit with mini1013/master_domain results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: metric value: 0.6511206701381028 name: Metric --- # SetFit with mini1013/master_domain This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - Model Type: SetFit - Sentence Transformer body: [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) - Classification head: a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - Maximum Sequence Length: 512 tokens - Number of Classes: 10 classes <!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Repository: [SetFit on GitHub](https://github.com/huggingface/setfit) - Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels \| Label \| Examples \| \|:------\|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| 9.0 \| <ul><li>'로저비비에 로저 비비어 i 러브 비비어 슬링백 펌프스 RVW53834670PE5 여성 37 주식회사 페칭'</li><li>'크롬베즈 스티치 장식 통굽펌프스 KP55797MA 카멜／245 sellerhub'</li><li>'HOBOKEN PS1511 PH2208 (3컬러) 브라운 230 NC_백화점'</li></ul> \| \| 2.0 \| <ul><li>'어그클래식울트라미니 ugg 어그부츠 여성 방한화 여자 발편한 겨울 신발 1116109 Sage Blossom_US 6(230) 울바이울'</li><li>'해외문스타 810s ET027 마르케 모디 운동화 장화 레인부츠 일본 직구 300_코요테_모디ET027 뉴저지홀세일'</li><li>'무릎 위에 앉다 장화 롱부츠 굽이 거칠다 평평한 바닥 고통 라이더 부츠 블랙_225 ZHANG YOUHUA'</li></ul> \| \| 0.0 \| <ul><li>'단화 한복신발 여성 새 혼례 소프트 한복구두 전통 꽃신 자수 39_빅화이트백봉이는한사이즈크게찍으셨으면좋겠습 대복컴퍼니'</li><li>'한복구두 꽃신 양단 생활한복 키높이 단화 굽 빅사이즈 담그어 여름 터지는 구슬 화이트-3.5cm_41 대한민국 일등 상점'</li><li>'여자 키높이 신발 여성 꽃신 한복 구두 전통 계량한복 37_화이트12(지연) 유럽걸스'</li></ul> \| \| 4.0 \| <ul><li>'남여공용 청키 클로그 바운서 샌들 (3ASDCBC33) 블랙(50BKS)_240 '</li><li>'[포멜카멜레]쥬얼장식트위드샌들 3cm FJS1F1SS024 아이보리/255 에이케이에스앤디(주) AK플라자 평택점'</li><li>'[하프클럽/] 에끌라 투웨이 주얼 샌들 33.카멜/245mm 롯데아이몰'</li></ul> \| \| 8.0 \| <ul><li>'에스콰이아 여성 발편한 경량 세미 캐주얼 앵클 워커 부츠 3cm J278C 브라운_230 (주) 패션플러스'</li><li>'[제옥스](신세계강남점) 스페리카 EC7 여성 워커부츠-블랙 W1B6VDJ3W11 블랙_245(38) 주식회사 에스에스지닷컴'</li><li>'(신세계강남점)금강 랜드로바 경량 컴포트 여성 워커 부츠 LANBOC4107WK1 240 신세계백화점'</li></ul> \| \| 6.0 \| <ul><li>'10mm 2중바닥 실내 슬리퍼 병원 거실 호텔 실내화 슬리퍼-타올천_고급-C_검정 주식회사 하루이'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_255 소프달링'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_245 소프달링'</li></ul> \| \| 3.0 \| <ul><li>'지안비토로씨 여성 마고 미드 부티 GIA36T75BLU18A1A00 EU 38.5 봉쥬르유럽'</li><li>'모다아울렛 121507 여성 7cm 깔끔 스틸레토 부티 구두 블랙k040_250 ◈217326053◈ MODA아울렛'</li><li>'미들부츠 미들힐 봄신상 워커 롱부츠 봄 가을신상 힐 블랙 245 바이포비'</li></ul> \| \| 5.0 \| <ul><li>'[공식판매] 버켄스탁 지제 에바 EVA 블랙 화이트 07 비트루트퍼플 키즈_220 (34) 좁은발볼 (Narrow) '</li><li>'eva 털슬리퍼 방한 방수 따듯한 털신 통굽 실내 화 기모 크로스오버 블랙M 소보로샵'</li><li>'크록스호환내피 털 탈부착 퍼 겨울 슬리퍼 안감 크림화이트(주니어)_C10-165(155~165) 인터코리아'</li></ul> \| \| 7.0 \| <ul><li>'[밸롭] 구름 브리즈 베이지 구름 브리즈 베이지245 (주)지티에스글로벌'</li><li>'[스텝100] 무지외반증 허리디스크 평발 신발 무릎 관절 중년 여성 운동화 화이트핑크플라워_235 스텝100'</li><li>'물컹슈즈 2.0 기능성 운동화 발편한 쿠션 운동화 무지외반증신발 족저근막염 물컹 업그레이드2.0_네이비_46(280mm) 주식회사 나인투식스'</li></ul> \| \| 1.0 \| <ul><li>'베라왕 스타일온에어 23SS 청 플랫폼 로퍼 80111682 G 667381 틸블루_230 DM ENG'</li><li>'[MUJI] 발수 발이 편한 스니커 머스터드 235mm 4550182676303 무인양품(주)'</li><li>'[반스(슈즈)]반스 어센틱 체커보드 스니커즈 (VN000W4NDI0) 4.240 롯데아이몰'</li></ul> \| ## Evaluation ### Metrics \| Label \| Metric \| \|:--------\|:-------\| \| all \| 0.6511 \| ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("mini1013/master_cate_ac10") # Run inference preds = model("XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다") ``` <!-- ### Downstream Use List how someone could finetune this model on their own dataset. --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Set Metrics \| Training set \| Min \| Median \| Max \| \|:-------------\|:----\|:-------\|:----\| \| Word count \| 3 \| 10.504 \| 21 \| \| Label \| Training Sample Count \| \|:------\|:----------------------\| \| 0.0 \| 50 \| \| 1.0 \| 50 \| \| 2.0 \| 50 \| \| 3.0 \| 50 \| \| 4.0 \| 50 \| \| 5.0 \| 50 \| \| 6.0 \| 50 \| \| 7.0 \| 50 \| \| 8.0 \| 50 \| \| 9.0 \| 50 \| ### Training Hyperparameters - batch_size: (512, 512) - num_epochs: (20, 20) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 40 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results \| Epoch \| Step \| Training Loss \| Validation Loss \| \|:-------:\|:----:\|:-------------:\|:---------------:\| \| 0.0127 \| 1 \| 0.4172 \| - \| \| 0.6329 \| 50 \| 0.3266 \| - \| \| 1.2658 \| 100 \| 0.1718 \| - \| \| 1.8987 \| 150 \| 0.095 \| - \| \| 2.5316 \| 200 \| 0.0257 \| - \| \| 3.1646 \| 250 \| 0.0142 \| - \| \| 3.7975 \| 300 \| 0.0026 \| - \| \| 4.4304 \| 350 \| 0.0164 \| - \| \| 5.0633 \| 400 \| 0.01 \| - \| \| 5.6962 \| 450 \| 0.0004 \| - \| \| 6.3291 \| 500 \| 0.0003 \| - \| \| 6.9620 \| 550 \| 0.0002 \| - \| \| 7.5949 \| 600 \| 0.0002 \| - \| \| 8.2278 \| 650 \| 0.0001 \| - \| \| 8.8608 \| 700 \| 0.0001 \| - \| \| 9.4937 \| 750 \| 0.0001 \| - \| \| 10.1266 \| 800 \| 0.0001 \| - \| \| 10.7595 \| 850 \| 0.0001 \| - \| \| 11.3924 \| 900 \| 0.0001 \| - \| \| 12.0253 \| 950 \| 0.0001 \| - \| \| 12.6582 \| 1000 \| 0.0001 \| - \| \| 13.2911 \| 1050 \| 0.0001 \| - \| \| 13.9241 \| 1100 \| 0.0001 \| - \| \| 14.5570 \| 1150 \| 0.0001 \| - \| \| 15.1899 \| 1200 \| 0.0001 \| - \| \| 15.8228 \| 1250 \| 0.0001 \| - \| \| 16.4557 \| 1300 \| 0.0001 \| - \| \| 17.0886 \| 1350 \| 0.0001 \| - \| \| 17.7215 \| 1400 \| 0.0001 \| - \| \| 18.3544 \| 1450 \| 0.0001 \| - \| \| 18.9873 \| 1500 \| 0.0001 \| - \| \| 19.6203 \| 1550 \| 0.0001 \| - \| ### Framework Versions - Python: 3.10.12 - SetFit: 1.1.0.dev0 - Sentence Transformers: 3.1.1 - Transformers: 4.46.1 - PyTorch: 2.4.0+cu121 - Datasets: 2.20.0 - Tokenizers: 0.20.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# SetFit with mini1013/master_domain This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - Model Type: SetFit - Sentence Transformer body: [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) - Classification head: a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - Maximum Sequence Length: 512 tokens - Number of Classes: 10 classes <!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Repository: [SetFit on GitHub](https://github.com/huggingface/setfit) - Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels \| Label \| Examples \| \|:------\|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| 9.0 \| <ul><li>'로저비비에 로저 비비어 i 러브 비비어 슬링백 펌프스 RVW53834670PE5 여성 37 주식회사 페칭'</li><li>'크롬베즈 스티치 장식 통굽펌프스 KP55797MA 카멜／245 sellerhub'</li><li>'HOBOKEN PS1511 PH2208 (3컬러) 브라운 230 NC_백화점'</li></ul> \| \| 2.0 \| <ul><li>'어그클래식울트라미니 ugg 어그부츠 여성 방한화 여자 발편한 겨울 신발 1116109 Sage Blossom_US 6(230) 울바이울'</li><li>'해외문스타 810s ET027 마르케 모디 운동화 장화 레인부츠 일본 직구 300_코요테_모디ET027 뉴저지홀세일'</li><li>'무릎 위에 앉다 장화 롱부츠 굽이 거칠다 평평한 바닥 고통 라이더 부츠 블랙_225 ZHANG YOUHUA'</li></ul> \| \| 0.0 \| <ul><li>'단화 한복신발 여성 새 혼례 소프트 한복구두 전통 꽃신 자수 39_빅화이트백봉이는한사이즈크게찍으셨으면좋겠습 대복컴퍼니'</li><li>'한복구두 꽃신 양단 생활한복 키높이 단화 굽 빅사이즈 담그어 여름 터지는 구슬 화이트-3.5cm_41 대한민국 일등 상점'</li><li>'여자 키높이 신발 여성 꽃신 한복 구두 전통 계량한복 37_화이트12(지연) 유럽걸스'</li></ul> \| \| 4.0 \| <ul><li>'남여공용 청키 클로그 바운서 샌들 (3ASDCBC33) 블랙(50BKS)_240 '</li><li>'[포멜카멜레]쥬얼장식트위드샌들 3cm FJS1F1SS024 아이보리/255 에이케이에스앤디(주) AK플라자 평택점'</li><li>'[하프클럽/] 에끌라 투웨이 주얼 샌들 33.카멜/245mm 롯데아이몰'</li></ul> \| \| 8.0 \| <ul><li>'에스콰이아 여성 발편한 경량 세미 캐주얼 앵클 워커 부츠 3cm J278C 브라운_230 (주) 패션플러스'</li><li>'[제옥스](신세계강남점) 스페리카 EC7 여성 워커부츠-블랙 W1B6VDJ3W11 블랙_245(38) 주식회사 에스에스지닷컴'</li><li>'(신세계강남점)금강 랜드로바 경량 컴포트 여성 워커 부츠 LANBOC4107WK1 240 신세계백화점'</li></ul> \| \| 6.0 \| <ul><li>'10mm 2중바닥 실내 슬리퍼 병원 거실 호텔 실내화 슬리퍼-타올천_고급-C_검정 주식회사 하루이'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_255 소프달링'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_245 소프달링'</li></ul> \| \| 3.0 \| <ul><li>'지안비토로씨 여성 마고 미드 부티 GIA36T75BLU18A1A00 EU 38.5 봉쥬르유럽'</li><li>'모다아울렛 121507 여성 7cm 깔끔 스틸레토 부티 구두 블랙k040_250 ◈217326053◈ MODA아울렛'</li><li>'미들부츠 미들힐 봄신상 워커 롱부츠 봄 가을신상 힐 블랙 245 바이포비'</li></ul> \| \| 5.0 \| <ul><li>'[공식판매] 버켄스탁 지제 에바 EVA 블랙 화이트 07 비트루트퍼플 키즈_220 (34) 좁은발볼 (Narrow) '</li><li>'eva 털슬리퍼 방한 방수 따듯한 털신 통굽 실내 화 기모 크로스오버 블랙M 소보로샵'</li><li>'크록스호환내피 털 탈부착 퍼 겨울 슬리퍼 안감 크림화이트(주니어)_C10-165(155~165) 인터코리아'</li></ul> \| \| 7.0 \| <ul><li>'[밸롭] 구름 브리즈 베이지 구름 브리즈 베이지245 (주)지티에스글로벌'</li><li>'[스텝100] 무지외반증 허리디스크 평발 신발 무릎 관절 중년 여성 운동화 화이트핑크플라워_235 스텝100'</li><li>'물컹슈즈 2.0 기능성 운동화 발편한 쿠션 운동화 무지외반증신발 족저근막염 물컹 업그레이드2.0_네이비_46(280mm) 주식회사 나인투식스'</li></ul> \| \| 1.0 \| <ul><li>'베라왕 스타일온에어 23SS 청 플랫폼 로퍼 80111682 G 667381 틸블루_230 DM ENG'</li><li>'[MUJI] 발수 발이 편한 스니커 머스터드 235mm 4550182676303 무인양품(주)'</li><li>'[반스(슈즈)]반스 어센틱 체커보드 스니커즈 (VN000W4NDI0) 4.240 롯데아이몰'</li></ul> \| ## Evaluation ### Metrics \| Label \| Metric \| \|:--------\|:-------\| \| all \| 0.6511 \| ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("mini1013/master_cate_ac10") # Run inference preds = model("XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다") ``` <!-- ### Downstream Use List how someone could finetune this model on their own dataset. --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Set Metrics \| Training set \| Min \| Median \| Max \| \|:-------------\|:----\|:-------\|:----\| \| Word count \| 3 \| 10.504 \| 21 \| \| Label \| Training Sample Count \| \|:------\|:----------------------\| \| 0.0 \| 50 \| \| 1.0 \| 50 \| \| 2.0 \| 50 \| \| 3.0 \| 50 \| \| 4.0 \| 50 \| \| 5.0 \| 50 \| \| 6.0 \| 50 \| \| 7.0 \| 50 \| \| 8.0 \| 50 \| \| 9.0 \| 50 \| ### Training Hyperparameters - batch_size: (512, 512) - num_epochs: (20, 20) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 40 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results \| Epoch \| Step \| Training Loss \| Validation Loss \| \|:-------:\|:----:\|:-------------:\|:---------------:\| \| 0.0127 \| 1 \| 0.4172 \| - \| \| 0.6329 \| 50 \| 0.3266 \| - \| \| 1.2658 \| 100 \| 0.1718 \| - \| \| 1.8987 \| 150 \| 0.095 \| - \| \| 2.5316 \| 200 \| 0.0257 \| - \| \| 3.1646 \| 250 \| 0.0142 \| - \| \| 3.7975 \| 300 \| 0.0026 \| - \| \| 4.4304 \| 350 \| 0.0164 \| - \| \| 5.0633 \| 400 \| 0.01 \| - \| \| 5.6962 \| 450 \| 0.0004 \| - \| \| 6.3291 \| 500 \| 0.0003 \| - \| \| 6.9620 \| 550 \| 0.0002 \| - \| \| 7.5949 \| 600 \| 0.0002 \| - \| \| 8.2278 \| 650 \| 0.0001 \| - \| \| 8.8608 \| 700 \| 0.0001 \| - \| \| 9.4937 \| 750 \| 0.0001 \| - \| \| 10.1266 \| 800 \| 0.0001 \| - \| \| 10.7595 \| 850 \| 0.0001 \| - \| \| 11.3924 \| 900 \| 0.0001 \| - \| \| 12.0253 \| 950 \| 0.0001 \| - \| \| 12.6582 \| 1000 \| 0.0001 \| - \| \| 13.2911 \| 1050 \| 0.0001 \| - \| \| 13.9241 \| 1100 \| 0.0001 \| - \| \| 14.5570 \| 1150 \| 0.0001 \| - \| \| 15.1899 \| 1200 \| 0.0001 \| - \| \| 15.8228 \| 1250 \| 0.0001 \| - \| \| 16.4557 \| 1300 \| 0.0001 \| - \| \| 17.0886 \| 1350 \| 0.0001 \| - \| \| 17.7215 \| 1400 \| 0.0001 \| - \| \| 18.3544 \| 1450 \| 0.0001 \| - \| \| 18.9873 \| 1500 \| 0.0001 \| - \| \| 19.6203 \| 1550 \| 0.0001 \| - \| ### Framework Versions - Python: 3.10.12 - SetFit: 1.1.0.dev0 - Sentence Transformers: 3.1.1 - Transformers: 4.46.1 - PyTorch: 2.4.0+cu121 - Datasets: 2.20.0 - Tokenizers: 0.20.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "mini1013/master_domain", "library_name": "setfit", "metrics": ["metric"], "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification", "generated_from_setfit_trainer"], "widget": [{"text": "벤시몽 RAIN BOOTS MID - 7color DOLPHIN GREY_40 260 오리상점"}, {"text": "플레이볼 오리진 뮬 (PLAYBALL ORIGIN MULE) NY (Off White) 화이트_230 주식회사 에프앤에프"}, {"text": "XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다"}, {"text": "다이어트 슬리퍼 다리 부종 스트레칭 균형 실내화 핑크 33-37_33 글로벌다이렉트"}, {"text": "케즈 챔피온 스트랩 캔버스5 M01778F001 Black/Black/Black_230 블루빌리"}], "inference": true, "model-index": [{"name": "SetFit with mini1013/master_domain", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "Unknown", "type": "unknown", "split": "test"}, "metrics": [{"type": "metric", "value": 0.6511206701381028, "name": "Metric"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,594
RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf	RichardErkhov	null	[ "gguf", "endpoints_compatible", "region:us", "conversational" ]	2024-10-30T02:40:41Z	2024-10-30T04:00:56+00:00	24	0	--- {} --- Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) gemma-2-2b-finance-it-v1 - GGUF - Model creator: https://huggingface.co/miner41612/ - Original model: https://huggingface.co/miner41612/gemma-2-2b-finance-it-v1/ \| Name \| Quant method \| Size \| \| ---- \| ---- \| ---- \| \| [gemma-2-2b-finance-it-v1.Q2_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q2_K.gguf) \| Q2_K \| 1.15GB \| \| [gemma-2-2b-finance-it-v1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_S.gguf) \| Q3_K_S \| 1.27GB \| \| [gemma-2-2b-finance-it-v1.Q3_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K.gguf) \| Q3_K \| 1.36GB \| \| [gemma-2-2b-finance-it-v1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_M.gguf) \| Q3_K_M \| 1.36GB \| \| [gemma-2-2b-finance-it-v1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_L.gguf) \| Q3_K_L \| 1.44GB \| \| [gemma-2-2b-finance-it-v1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_XS.gguf) \| IQ4_XS \| 1.47GB \| \| [gemma-2-2b-finance-it-v1.Q4_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_0.gguf) \| Q4_0 \| 1.52GB \| \| [gemma-2-2b-finance-it-v1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_NL.gguf) \| IQ4_NL \| 1.53GB \| \| [gemma-2-2b-finance-it-v1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_S.gguf) \| Q4_K_S \| 1.53GB \| \| [gemma-2-2b-finance-it-v1.Q4_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K.gguf) \| Q4_K \| 1.59GB \| \| [gemma-2-2b-finance-it-v1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_M.gguf) \| Q4_K_M \| 1.59GB \| \| [gemma-2-2b-finance-it-v1.Q4_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_1.gguf) \| Q4_1 \| 1.64GB \| \| [gemma-2-2b-finance-it-v1.Q5_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_0.gguf) \| Q5_0 \| 1.75GB \| \| [gemma-2-2b-finance-it-v1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_S.gguf) \| Q5_K_S \| 1.75GB \| \| [gemma-2-2b-finance-it-v1.Q5_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K.gguf) \| Q5_K \| 1.79GB \| \| [gemma-2-2b-finance-it-v1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_M.gguf) \| Q5_K_M \| 1.79GB \| \| [gemma-2-2b-finance-it-v1.Q5_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_1.gguf) \| Q5_1 \| 1.87GB \| \| [gemma-2-2b-finance-it-v1.Q6_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q6_K.gguf) \| Q6_K \| 2.0GB \| \| [gemma-2-2b-finance-it-v1.Q8_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q8_0.gguf) \| Q8_0 \| 2.59GB \| Original model description: --- base_model: - miner41612/gemma-2-2b-finance-it-v1 datasets: - Mineru/kor-open-finance - Mineru/kor-finance-sft language: - ko library_name: transformers license: gemma pipeline_tag: text-generation tags: - krx - finance - sft - trl extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- # Gemma 2 Finance model card Model Page: [Gemma](https://ai.google.dev/gemma/docs/base) Terms of Use: [Terms][terms] Authors: miner41612 ## Model Information 입력 및 출력에 대한 요약 설명과 간략한 정의입니다. ### Description Google의 Gemma 2 2b 모델을 금융 도메인 데이터셋을 정재한 데이터셋을 Continual Learning을 하여 학습 한 모델에 금융 도메인 Insturction 데이터 셋으로 학습 시킨 모델입니다. ### Usage Below we share some code snippets on how to get quickly started with running the model. First, install the Transformers library with: ```sh pip install -U transformers ``` Then, copy the snippet from the section that is relevant for your usecase. #### Running with the `pipeline` API ```python import torch from transformers import pipeline pipe = pipeline( "text-generation", model="miner41612/gemma-2-2b-finance-it-v1", model_kwargs={"torch_dtype": torch.bfloat16}, device="cuda", # replace with "mps" to run on a Mac device ) messages = [ {"role": "user", "content": "원가상환제도란?"}, ] outputs = pipe(messages, max_new_tokens=256) assistant_response = outputs[0]["generated_text"][-1]["content"].strip() print(assistant_response) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", device_map="auto", ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` You can ensure the correct chat template is applied by using `tokenizer.apply_chat_template` as follows: ```python messages = [ {"role": "user", "content": "원가상환제도란?"}, ] input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True).to("cuda") outputs = model.generate(input_ids, max_new_tokens=256) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` <details> <summary> Using 8-bit precision (int8) </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> <details> <summary> Using 4-bit precision </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> #### Advanced Usage <details> <summary> Torch compile </summary> [Torch compile](https://pytorch.org/tutorials/intermediate/torch_compile_tutorial.html) is a method for speeding-up the inference of PyTorch modules. The Gemma-2 2b model can be run up to 6x faster by leveraging torch compile. Note that two warm-up steps are required before the full inference speed is realised: ```python import os os.environ["TOKENIZERS_PARALLELISM"] = "false" from transformers import AutoTokenizer, Gemma2ForCausalLM from transformers.cache_utils import HybridCache import torch torch.set_float32_matmul_precision("high") # load the model + tokenizer tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = Gemma2ForCausalLM.from_pretrained("miner41612/gemma-2-2b-finance-it-v1", torch_dtype=torch.bfloat16) model.to("cuda") # apply the torch compile transformation model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) # pre-process inputs input_text = "원가상환제도란? " model_inputs = tokenizer(input_text, return_tensors="pt").to("cuda") prompt_length = model_inputs.input_ids.shape[1] # set-up k/v cache past_key_values = HybridCache( config=model.config, max_batch_size=1, max_cache_len=model.config.max_position_embeddings, device=model.device, dtype=model.dtype ) # enable passing kv cache to generate model._supports_cache_class = True model.generation_config.cache_implementation = None # two warm-up steps for idx in range(2): outputs = model.generate(model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) past_key_values.reset() # fast run outputs = model.generate(model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` For more details, refer to the [Transformers documentation](https://huggingface.co/docs/transformers/main/en/llm_optims?static-kv=basic+usage%3A+generation_config). </details> ### Inputs and outputs * Input: Text string, such as a question, a prompt, or a document to be summarized. * Output: Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ### Citation ```none @article{gemma_2024, title={Gemma}, url={https://www.kaggle.com/m/3301}, DOI={10.34740/KAGGLE/M/3301}, publisher={Kaggle}, author={Gemma Team}, year={2024} } ``` ## Model Data Data used for model training and how the data was processed. ## Ethics and Safety Ethics and safety evaluation approach and results. ## Dangerous Capability Evaluations ### Evaluation Approach We evaluated a range of dangerous capabilities: - Offensive cybersecurity: To assess the model's potential for misuse in cybersecurity contexts, we utilized both publicly available Capture-the-Flag (CTF) platforms like InterCode-CTF and Hack the Box, as well as internally developed CTF challenges. These evaluations measure the model's ability to exploit vulnerabilities and gain unauthorized access in simulated environments. - Self-proliferation: We evaluated the model's capacity for self-proliferation by designing tasks that involve resource acquisition, code execution, and interaction with remote systems. These evaluations assess the model's ability to independently replicate and spread. - Persuasion: To evaluate the model's capacity for persuasion and deception, we conducted human persuasion studies. These studies involved scenarios that measure the model's ability to build rapport, influence beliefs, and elicit specific actions from human participants. ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit][rai-toolkit]. * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy][prohibited-use]. * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques.	null	Non_BioNLP	Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) gemma-2-2b-finance-it-v1 - GGUF - Model creator: https://huggingface.co/miner41612/ - Original model: https://huggingface.co/miner41612/gemma-2-2b-finance-it-v1/ \| Name \| Quant method \| Size \| \| ---- \| ---- \| ---- \| \| [gemma-2-2b-finance-it-v1.Q2_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q2_K.gguf) \| Q2_K \| 1.15GB \| \| [gemma-2-2b-finance-it-v1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_S.gguf) \| Q3_K_S \| 1.27GB \| \| [gemma-2-2b-finance-it-v1.Q3_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K.gguf) \| Q3_K \| 1.36GB \| \| [gemma-2-2b-finance-it-v1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_M.gguf) \| Q3_K_M \| 1.36GB \| \| [gemma-2-2b-finance-it-v1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_L.gguf) \| Q3_K_L \| 1.44GB \| \| [gemma-2-2b-finance-it-v1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_XS.gguf) \| IQ4_XS \| 1.47GB \| \| [gemma-2-2b-finance-it-v1.Q4_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_0.gguf) \| Q4_0 \| 1.52GB \| \| [gemma-2-2b-finance-it-v1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_NL.gguf) \| IQ4_NL \| 1.53GB \| \| [gemma-2-2b-finance-it-v1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_S.gguf) \| Q4_K_S \| 1.53GB \| \| [gemma-2-2b-finance-it-v1.Q4_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K.gguf) \| Q4_K \| 1.59GB \| \| [gemma-2-2b-finance-it-v1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_M.gguf) \| Q4_K_M \| 1.59GB \| \| [gemma-2-2b-finance-it-v1.Q4_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_1.gguf) \| Q4_1 \| 1.64GB \| \| [gemma-2-2b-finance-it-v1.Q5_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_0.gguf) \| Q5_0 \| 1.75GB \| \| [gemma-2-2b-finance-it-v1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_S.gguf) \| Q5_K_S \| 1.75GB \| \| [gemma-2-2b-finance-it-v1.Q5_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K.gguf) \| Q5_K \| 1.79GB \| \| [gemma-2-2b-finance-it-v1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_M.gguf) \| Q5_K_M \| 1.79GB \| \| [gemma-2-2b-finance-it-v1.Q5_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_1.gguf) \| Q5_1 \| 1.87GB \| \| [gemma-2-2b-finance-it-v1.Q6_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q6_K.gguf) \| Q6_K \| 2.0GB \| \| [gemma-2-2b-finance-it-v1.Q8_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q8_0.gguf) \| Q8_0 \| 2.59GB \| Original model description: --- base_model: - miner41612/gemma-2-2b-finance-it-v1 datasets: - Mineru/kor-open-finance - Mineru/kor-finance-sft language: - ko library_name: transformers license: gemma pipeline_tag: text-generation tags: - krx - finance - sft - trl extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- # Gemma 2 Finance model card Model Page: [Gemma](https://ai.google.dev/gemma/docs/base) Terms of Use: [Terms][terms] Authors: miner41612 ## Model Information 입력 및 출력에 대한 요약 설명과 간략한 정의입니다. ### Description Google의 Gemma 2 2b 모델을 금융 도메인 데이터셋을 정재한 데이터셋을 Continual Learning을 하여 학습 한 모델에 금융 도메인 Insturction 데이터 셋으로 학습 시킨 모델입니다. ### Usage Below we share some code snippets on how to get quickly started with running the model. First, install the Transformers library with: ```sh pip install -U transformers ``` Then, copy the snippet from the section that is relevant for your usecase. #### Running with the `pipeline` API ```python import torch from transformers import pipeline pipe = pipeline( "text-generation", model="miner41612/gemma-2-2b-finance-it-v1", model_kwargs={"torch_dtype": torch.bfloat16}, device="cuda", # replace with "mps" to run on a Mac device ) messages = [ {"role": "user", "content": "원가상환제도란?"}, ] outputs = pipe(messages, max_new_tokens=256) assistant_response = outputs[0]["generated_text"][-1]["content"].strip() print(assistant_response) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", device_map="auto", ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` You can ensure the correct chat template is applied by using `tokenizer.apply_chat_template` as follows: ```python messages = [ {"role": "user", "content": "원가상환제도란?"}, ] input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True).to("cuda") outputs = model.generate(input_ids, max_new_tokens=256) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` <details> <summary> Using 8-bit precision (int8) </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> <details> <summary> Using 4-bit precision </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> #### Advanced Usage <details> <summary> Torch compile </summary> [Torch compile](https://pytorch.org/tutorials/intermediate/torch_compile_tutorial.html) is a method for speeding-up the inference of PyTorch modules. The Gemma-2 2b model can be run up to 6x faster by leveraging torch compile. Note that two warm-up steps are required before the full inference speed is realised: ```python import os os.environ["TOKENIZERS_PARALLELISM"] = "false" from transformers import AutoTokenizer, Gemma2ForCausalLM from transformers.cache_utils import HybridCache import torch torch.set_float32_matmul_precision("high") # load the model + tokenizer tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = Gemma2ForCausalLM.from_pretrained("miner41612/gemma-2-2b-finance-it-v1", torch_dtype=torch.bfloat16) model.to("cuda") # apply the torch compile transformation model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) # pre-process inputs input_text = "원가상환제도란? " model_inputs = tokenizer(input_text, return_tensors="pt").to("cuda") prompt_length = model_inputs.input_ids.shape[1] # set-up k/v cache past_key_values = HybridCache( config=model.config, max_batch_size=1, max_cache_len=model.config.max_position_embeddings, device=model.device, dtype=model.dtype ) # enable passing kv cache to generate model._supports_cache_class = True model.generation_config.cache_implementation = None # two warm-up steps for idx in range(2): outputs = model.generate(model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) past_key_values.reset() # fast run outputs = model.generate(model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` For more details, refer to the [Transformers documentation](https://huggingface.co/docs/transformers/main/en/llm_optims?static-kv=basic+usage%3A+generation_config). </details> ### Inputs and outputs * Input: Text string, such as a question, a prompt, or a document to be summarized. * Output: Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ### Citation ```none @article{gemma_2024, title={Gemma}, url={https://www.kaggle.com/m/3301}, DOI={10.34740/KAGGLE/M/3301}, publisher={Kaggle}, author={Gemma Team}, year={2024} } ``` ## Model Data Data used for model training and how the data was processed. ## Ethics and Safety Ethics and safety evaluation approach and results. ## Dangerous Capability Evaluations ### Evaluation Approach We evaluated a range of dangerous capabilities: - Offensive cybersecurity: To assess the model's potential for misuse in cybersecurity contexts, we utilized both publicly available Capture-the-Flag (CTF) platforms like InterCode-CTF and Hack the Box, as well as internally developed CTF challenges. These evaluations measure the model's ability to exploit vulnerabilities and gain unauthorized access in simulated environments. - Self-proliferation: We evaluated the model's capacity for self-proliferation by designing tasks that involve resource acquisition, code execution, and interaction with remote systems. These evaluations assess the model's ability to independently replicate and spread. - Persuasion: To evaluate the model's capacity for persuasion and deception, we conducted human persuasion studies. These studies involved scenarios that measure the model's ability to build rapport, influence beliefs, and elicit specific actions from human participants. ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit][rai-toolkit]. * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy][prohibited-use]. * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques.	{}	task	[ "SUMMARIZATION" ]	44,595
9unu/formal_speech_translation	9unu	text2text-generation	[ "transformers", "safetensors", "t5", "text2text-generation", "license:mit", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2024-05-13T10:07:45Z	2024-05-27T10:12:59+00:00	4	0	--- license: mit --- # formal_speech_translation 카카오톡 AI 말투 변환기의 역공학 엔지니어링을 통해 개발한 대화체에 견고한 상냥체 변환기입니다. [t5모델](https://huggingface.co/google-t5)을 한국어 전용 데이터로 학습한 [pko-t5-base](https://huggingface.co/paust/pko-t5-base) 모델을 활용하여 학습을 진행했습니다. [dev_repository](https://github.com/9unu/Kakao_Reverse-Engineering) ## Usage transformers의 pipe, from_pretrained 등의 api를 활용하여 접근 가능합니다. ## Example ```python from transformers import T5TokenizerFast, T5ForConditionalGeneration, pipeline # 모델 경로 및 cpu, gpu 지정 cache_dir = "./hugging_face" gentle_model_path='9unu/formal_speech_translation' gentle_model = T5ForConditionalGeneration.from_pretrained(formal_model_path, cache_dir=cache_dir) device = torch.device("cuda" if torch.cuda.is_available() else "cpu") ``` ```python # transformers 파이프라인 생성 gentle_pipeline = pipeline(model = gentle_model, tokenizer = tokenizer, device = device, max_length=60) # text 말투 변환 text = "밥 먹는 중이야" num_return_sequences = 1 max_length = 60 out = gentle_pipeline(text, num_return_sequences = num_return_sequences, max_length=max_length) print([x['generated_text'] for x in out]) ``` ## License 본 모델은 MIT license 하에 공개되어 있습니다.	null	Non_BioNLP	# formal_speech_translation 카카오톡 AI 말투 변환기의 역공학 엔지니어링을 통해 개발한 대화체에 견고한 상냥체 변환기입니다. [t5모델](https://huggingface.co/google-t5)을 한국어 전용 데이터로 학습한 [pko-t5-base](https://huggingface.co/paust/pko-t5-base) 모델을 활용하여 학습을 진행했습니다. [dev_repository](https://github.com/9unu/Kakao_Reverse-Engineering) ## Usage transformers의 pipe, from_pretrained 등의 api를 활용하여 접근 가능합니다. ## Example ```python from transformers import T5TokenizerFast, T5ForConditionalGeneration, pipeline # 모델 경로 및 cpu, gpu 지정 cache_dir = "./hugging_face" gentle_model_path='9unu/formal_speech_translation' gentle_model = T5ForConditionalGeneration.from_pretrained(formal_model_path, cache_dir=cache_dir) device = torch.device("cuda" if torch.cuda.is_available() else "cpu") ``` ```python # transformers 파이프라인 생성 gentle_pipeline = pipeline(model = gentle_model, tokenizer = tokenizer, device = device, max_length=60) # text 말투 변환 text = "밥 먹는 중이야" num_return_sequences = 1 max_length = 60 out = gentle_pipeline(text, num_return_sequences = num_return_sequences, max_length=max_length) print([x['generated_text'] for x in out]) ``` ## License 본 모델은 MIT license 하에 공개되어 있습니다.	{"license": "mit"}	task	[ "TRANSLATION" ]	44,596
somosnlp/gemma-7b-it-legal-refugiados-es	somosnlp	null	[ "transformers, pe", "safetensors", "gemma", "trl", "sft", "generated_from_trainer", "es", "dataset:somosnlp/instruct-legal-refugiados-es", "arxiv:1910.09700", "base_model:google/gemma-7b", "base_model:finetune:google/gemma-7b", "license:apache-2.0", "region:us" ]	2024-03-18T22:59:55Z	2024-04-24T11:36:00+00:00	16	0	--- base_model: google/gemma-7b datasets: - somosnlp/instruct-legal-refugiados-es language: - es library_name: transformers, pe license: apache-2.0 tags: - trl - sft - generated_from_trainer --- <!-- Esta plantilla de Model Card es una adaptación de la de Hugging Face: https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md ¿Cómo utilizar esta plantilla? Copia el contenido en el README.md del repo de tu modelo en el Hub de Hugging Face y rellena cada sección. Para más información sobre cómo rellenar cada sección ver las docs: https://huggingface.co/docs/hub/model-cards --> # Model Card for gemma-7b-it-legal-refugiados-es <!-- Suele haber un nombre corto ("pretty name") para las URLs, tablas y demás y uno largo más descriptivo. Para crear el pretty name podéis utilizar acrónimos. --> <!-- Resumen del modelo y motivación del proyecto (inc. los ODS relacionados). Esta sección es como el abstract. También se puede incluir aquí el logo del proyecto. --> <!-- Si queréis incluir una versión de la Dataset Card en español, enlazarla aquí al principio (e.g. `README_es.md`).--> Spain is the third country with the highest number of asylum applications, receiving each year approximately more than 100,000 applications, and the third with the lowest number of approvals within the EU. The main objective of this project is to facilitate the tasks of NGOs in this field and other institutions and help them to obtain answers to questions (QA) related to refugee legislation in Spanish. With its refined understanding of the nuances and complexities of this legal field. The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain ## Model Details ### Model Description <!-- Resumen del modelo. --> The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the dataset [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es). This is the model card of a 🤗 transformers model that has been pushed on the Hub to allow public access. - Developed by: <!-- Nombre de los miembros del equipo --> [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim) - Funded by: SomosNLP, HuggingFace <!-- Si contasteis con apoyo de otra entidad (e.g. vuestra universidad), añadidla aquí --> - Model type: Language model, instruction tuned - Language(s): es-ES, es-MX, es-VE <!-- Enumerar las lenguas en las que se ha entrenado el modelo, especificando el país de origen. Utilizar códigos ISO. Por ejemplo: Spanish (`es-CL`, `es-ES`, `es-MX`), Catalan (`ca`), Quechua (`qu`). --> - License: apache-2.0 <!-- Elegid una licencia lo más permisiva posible teniendo en cuenta la licencia del model pre-entrenado y los datasets utilizados --> - Fine-tuned from model: [google/gemma-7b](https://huggingface.co/google/ <!-- Enlace al modelo pre-entrenado que habéis utilizado como base --> - Dataset used: [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es) <!-- Enlace al dataset utilizado para el ajuste --> ### Model Sources - Repository: Notebook in [This repo](https://huggingface.co/somosnlp/gemma-7b-it-legal-refugee-v0.1.1) <!-- Enlace al `main` del repo donde tengáis los scripts, i.e.: o del mismo repo del modelo en HuggingFace o a GitHub. --> - Demo: [Demo Space](https://huggingface.co/spaces/somosnlp/QA-legal-refugiados) <!-- Enlace a la demo --> - Video presentation: [Youtube Video](https://www.youtube.com/watch?v=1OqHDE5LKMI&list=PLTA-KAy8nxaASMwEUWkkTfMaDxWBxn-8J&index=3) <!-- Enlace a vuestro vídeo de presentación en YouTube (están todos subidos aquí: https://www.youtube.com/playlist?list=PLTA-KAy8nxaASMwEUWkkTfMaDxWBxn-8J) --> ### Model Family <!-- Si habéis entrenado varios modelos similares podéis enumerarlos aquí. --> This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b). ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> The primary objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. ### Downstream Use <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> Intented to be use in question-answering with a context and text generation. ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> Misuse includes any application that promotes unethical practices, misinterprets refugee law, or uses the model for malicious purposes. The model is not designed to replace professional legal advice. ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> The model, while powerful, has limitations inherent to AI, including biases present in the training data. It may not cover all nuances of refugee regulations or adapt to changes in law without updates. ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> <!-- Example: Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. --> ## How to Get Started with the Model Use the code below to get started with the model. ```python import torch from transformers import ( AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, pipeline ) model_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer = AutoTokenizer.from_pretrained(tokenizer_id) # Cargamos el modelo en 4 bits para agilizar la inferencia quantization_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16, bnb_4bit_use_double_quant=True, ) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, device_map="auto", quantization_config=quantization_config, ) # Generamos el pipeline de generación de texto pipe = pipeline("text-generation", model=model, tokenizer=tokenizer) # Definimos el eos token para el modelo eos_token = tokenizer("<\|im_end\|>",add_special_tokens=False)["input_ids"][0] def generate_inference(instruction, input, temperature): prompt = pipe.tokenizer.apply_chat_template([{"role": "user", "content": f"{instruction}/n{input}"}], tokenize=False, add_generation_prompt=True) outputs = pipe(prompt, max_new_tokens=256, do_sample=True, num_beams=1, temperature=float(temperature), top_k=50, top_p=0.95, max_time= 300, eos_token_id=eos_token) return outputs[0]['generated_text'][len(prompt):].strip() instruction = "¿Podrías explicarme brevemente los hechos que originan el procedimiento y las posibles calificaciones, así como las sanciones correspondientes, según lo expuesto en el contexto?" input = "b) Hechos que motivan la incoación del procedimiento sucintamente expuestos, su posible calificación y las sanciones que pudieran corresponder, sin perjuicio de lo que resulte de la instrucción. c) Instructor y, en su caso, secretario del procedimiento, con expresa indicación del régimen de recusación de éstos. d) Órgano competente para la resolución del expediente y norma que le atribuye tal competencia. e) Indicación de la posibilidad de que el presunto responsable pueda reconocer voluntariamente su responsabilidad. f) Medidas de carácter provisional que se hayan acordado por el órgano competente para iniciar el procedimiento sancionador, sin perjuicio de las que se puedan adoptar durante éste de conformidad con los artículos 55 y 61 de la Ley Orgánica 4/2000, de 11 de enero. g) Indicación del derecho a formular alegaciones y a la audiencia en el procedimiento y de los plazos para su ejercicio. 2. El acuerdo de iniciación se comunicará al instructor con traslado de cuantas actuaciones existan al respecto y se notificará a los interesados, entendiéndose en todo caso por tal al expedientado. En la notificación se advertirá a los interesados que, de no efectuar alegaciones sobre el contenido de la iniciación del procedimiento en el plazo previsto en el artículo siguiente, no realizarse propuesta de prueba o no ser admitidas, por improcedentes o innecesarias, las pruebas propuestas, la iniciación podrá ser considerada propuesta de resolución cuando contenga un pronunciamiento preciso acerca de la responsabilidad imputada, con los efectos previstos en los artículos 229 y 230." response = test_inference(instruction, input, 0.3) print(f"Response:\n{response}") ``` ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> The dataset used was [instruct-legal-refugiados-es](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es) but we adapted the dataset to a ChatML format, described in the next section. ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> <!-- Detallar la técnica de entrenamiento utilizada y enlazar los scripts/notebooks. --> The training was done using RTX 4090 from Vast.ai with PeRF and Lora #### Preprocessing We wanted to make a conversation model so we investigated the base model prompt in order to make conversational base on [chatml format](https://github.com/MicrosoftDocs/azure-docs/blob/main/articles/ai-services/openai/includes/chat-markup-language.md#working-with-chat-markup-language-chatml) we identified the special tokens so the model could understand the different roles in the conversation Example ``` <bos><\|im_start\|>system You are Gemma.<\|im_end\|> <\|im_start\|>user Hello, how are you?<\|im_end\|> <\|im_start\|>assistant I'm doing great. How can I help you today?<\|im_end\|>\n<eos> ``` So we used [Phil Schmid's gemma chatml tokenizer](https://huggingface.co/philschmid/gemma-tokenizer-chatml) to adapt our dataset for training #### Training Hyperparameters <!-- Enumerar los valores de los hiperparámetros de entrenamiento. --> The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 66 - gradient_accumulation_steps: 2 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 3 - Training regime: <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results <!-- Enlazar aquí los scripts/notebooks de evaluación y especificar los resultados. --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly. --> <!-- Rellenar la información de la lista y calcular las emisiones con la página mencionada. --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: 1 X RTX4090 - Hours used: 4 - Cloud Provider: Vast.ai - Compute Region: West Europe - Carbon Emitted: 350W x 4h = 1.4 kWh x 0.57 kg eq. CO2/kWh = 0.8 kg eq. CO2 ## Technical Specifications <!-- Esta sección es opcional porque seguramente ya habéis mencionado estos detalles más arriba, igualmente está bien incluirlos aquí de nuevo como bullet points a modo de resumen. --> ### Model Architecture and Objective The base model is [google/gemma-7b](https://huggingface.co/google/gemma-7b) finetuned in 4-bit. ### Compute Infrastructure #### Hardware <!-- Indicar el hardware utilizado, podéis agradecer aquí a quien lo patrocinó. --> 1 x RTX4090 GPU by Vast.ai. #### Software <!-- Enumerar las librerías utilizadas (e.g. transformers, distilabel). --> Libraries: - transformers - bitsandbytes - accelerate - xformers - trl - peft - wandb ## License <!-- Indicar bajo qué licencia se libera el modelo explicando, si no es apache 2.0, a qué se debe la licencia más restrictiva (i.e. herencia de las licencias del modelo pre-entrenado o de los datos utilizados). --> This model is under the license of the Gemma models by Google. Link to consent: https://www.kaggle.com/models/google/gemma/license/consent ## Citation BibTeX: [More Information Needed] <!-- Aquí tenéis un ejemplo de cita de un dataset que podéis adaptar: ``` @software{benallal2024cosmopedia, author = {Ben Allal, Loubna and Lozhkov, Anton and Penedo, Guilherme and Wolf, Thomas and von Werra, Leandro}, title = {Cosmopedia}, month = February, year = 2024, url = {https://huggingface.co/datasets/HuggingFaceTB/cosmopedia} } ``` - benallal2024cosmopedia -> nombre + año + nombre del modelo - author: lista de miembros del equipo - title: nombre del modelo - year: año - url: enlace al modelo --> ``` @software{somosnlp2024asistenciarefugiados, author = {Alvaro Hidalgo, Eduardo Muñoz, Teresa Martín}, title = {gemma-7b-it-legal-refugiados-es}, month = April, year = 2024, url = {somosnlp/gemma-7b-it-legal-refugee-v0.1.1} } ``` ## More Information <!-- Indicar aquí que el marco en el que se desarrolló el proyecto, en esta sección podéis incluir agradecimientos y más información sobre los miembros del equipo. Podéis adaptar el ejemplo a vuestro gusto. --> This project was developed during the [Hackathon #Somos600M](https://somosnlp.org/hackathon) organized by SomosNLP. The model was trained using GPUs sponsored by HuggingFace. Team: [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim) <!-- - [Name 1](Link to Hugging Face profile) - [Name 2](Link to Hugging Face profile) --> ## Contact [optional] <!-- Email de contacto para´posibles preguntas sobre el modelo. -->	null	Non_BioNLP	<!-- Esta plantilla de Model Card es una adaptación de la de Hugging Face: https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md ¿Cómo utilizar esta plantilla? Copia el contenido en el README.md del repo de tu modelo en el Hub de Hugging Face y rellena cada sección. Para más información sobre cómo rellenar cada sección ver las docs: https://huggingface.co/docs/hub/model-cards --> # Model Card for gemma-7b-it-legal-refugiados-es <!-- Suele haber un nombre corto ("pretty name") para las URLs, tablas y demás y uno largo más descriptivo. Para crear el pretty name podéis utilizar acrónimos. --> <!-- Resumen del modelo y motivación del proyecto (inc. los ODS relacionados). Esta sección es como el abstract. También se puede incluir aquí el logo del proyecto. --> <!-- Si queréis incluir una versión de la Dataset Card en español, enlazarla aquí al principio (e.g. `README_es.md`).--> Spain is the third country with the highest number of asylum applications, receiving each year approximately more than 100,000 applications, and the third with the lowest number of approvals within the EU. The main objective of this project is to facilitate the tasks of NGOs in this field and other institutions and help them to obtain answers to questions (QA) related to refugee legislation in Spanish. With its refined understanding of the nuances and complexities of this legal field. The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain ## Model Details ### Model Description <!-- Resumen del modelo. --> The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the dataset [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es). This is the model card of a 🤗 transformers model that has been pushed on the Hub to allow public access. - Developed by: <!-- Nombre de los miembros del equipo --> [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim) - Funded by: SomosNLP, HuggingFace <!-- Si contasteis con apoyo de otra entidad (e.g. vuestra universidad), añadidla aquí --> - Model type: Language model, instruction tuned - Language(s): es-ES, es-MX, es-VE <!-- Enumerar las lenguas en las que se ha entrenado el modelo, especificando el país de origen. Utilizar códigos ISO. Por ejemplo: Spanish (`es-CL`, `es-ES`, `es-MX`), Catalan (`ca`), Quechua (`qu`). --> - License: apache-2.0 <!-- Elegid una licencia lo más permisiva posible teniendo en cuenta la licencia del model pre-entrenado y los datasets utilizados --> - Fine-tuned from model: [google/gemma-7b](https://huggingface.co/google/ <!-- Enlace al modelo pre-entrenado que habéis utilizado como base --> - Dataset used: [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es) <!-- Enlace al dataset utilizado para el ajuste --> ### Model Sources - Repository: Notebook in [This repo](https://huggingface.co/somosnlp/gemma-7b-it-legal-refugee-v0.1.1) <!-- Enlace al `main` del repo donde tengáis los scripts, i.e.: o del mismo repo del modelo en HuggingFace o a GitHub. --> - Demo: [Demo Space](https://huggingface.co/spaces/somosnlp/QA-legal-refugiados) <!-- Enlace a la demo --> - Video presentation: [Youtube Video](https://www.youtube.com/watch?v=1OqHDE5LKMI&list=PLTA-KAy8nxaASMwEUWkkTfMaDxWBxn-8J&index=3) <!-- Enlace a vuestro vídeo de presentación en YouTube (están todos subidos aquí: https://www.youtube.com/playlist?list=PLTA-KAy8nxaASMwEUWkkTfMaDxWBxn-8J) --> ### Model Family <!-- Si habéis entrenado varios modelos similares podéis enumerarlos aquí. --> This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b). ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> The primary objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. ### Downstream Use <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> Intented to be use in question-answering with a context and text generation. ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> Misuse includes any application that promotes unethical practices, misinterprets refugee law, or uses the model for malicious purposes. The model is not designed to replace professional legal advice. ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> The model, while powerful, has limitations inherent to AI, including biases present in the training data. It may not cover all nuances of refugee regulations or adapt to changes in law without updates. ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> <!-- Example: Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. --> ## How to Get Started with the Model Use the code below to get started with the model. ```python import torch from transformers import ( AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, pipeline ) model_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer = AutoTokenizer.from_pretrained(tokenizer_id) # Cargamos el modelo en 4 bits para agilizar la inferencia quantization_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16, bnb_4bit_use_double_quant=True, ) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, device_map="auto", quantization_config=quantization_config, ) # Generamos el pipeline de generación de texto pipe = pipeline("text-generation", model=model, tokenizer=tokenizer) # Definimos el eos token para el modelo eos_token = tokenizer("<\|im_end\|>",add_special_tokens=False)["input_ids"][0] def generate_inference(instruction, input, temperature): prompt = pipe.tokenizer.apply_chat_template([{"role": "user", "content": f"{instruction}/n{input}"}], tokenize=False, add_generation_prompt=True) outputs = pipe(prompt, max_new_tokens=256, do_sample=True, num_beams=1, temperature=float(temperature), top_k=50, top_p=0.95, max_time= 300, eos_token_id=eos_token) return outputs[0]['generated_text'][len(prompt):].strip() instruction = "¿Podrías explicarme brevemente los hechos que originan el procedimiento y las posibles calificaciones, así como las sanciones correspondientes, según lo expuesto en el contexto?" input = "b) Hechos que motivan la incoación del procedimiento sucintamente expuestos, su posible calificación y las sanciones que pudieran corresponder, sin perjuicio de lo que resulte de la instrucción. c) Instructor y, en su caso, secretario del procedimiento, con expresa indicación del régimen de recusación de éstos. d) Órgano competente para la resolución del expediente y norma que le atribuye tal competencia. e) Indicación de la posibilidad de que el presunto responsable pueda reconocer voluntariamente su responsabilidad. f) Medidas de carácter provisional que se hayan acordado por el órgano competente para iniciar el procedimiento sancionador, sin perjuicio de las que se puedan adoptar durante éste de conformidad con los artículos 55 y 61 de la Ley Orgánica 4/2000, de 11 de enero. g) Indicación del derecho a formular alegaciones y a la audiencia en el procedimiento y de los plazos para su ejercicio. 2. El acuerdo de iniciación se comunicará al instructor con traslado de cuantas actuaciones existan al respecto y se notificará a los interesados, entendiéndose en todo caso por tal al expedientado. En la notificación se advertirá a los interesados que, de no efectuar alegaciones sobre el contenido de la iniciación del procedimiento en el plazo previsto en el artículo siguiente, no realizarse propuesta de prueba o no ser admitidas, por improcedentes o innecesarias, las pruebas propuestas, la iniciación podrá ser considerada propuesta de resolución cuando contenga un pronunciamiento preciso acerca de la responsabilidad imputada, con los efectos previstos en los artículos 229 y 230." response = test_inference(instruction, input, 0.3) print(f"Response:\n{response}") ``` ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> The dataset used was [instruct-legal-refugiados-es](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es) but we adapted the dataset to a ChatML format, described in the next section. ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> <!-- Detallar la técnica de entrenamiento utilizada y enlazar los scripts/notebooks. --> The training was done using RTX 4090 from Vast.ai with PeRF and Lora #### Preprocessing We wanted to make a conversation model so we investigated the base model prompt in order to make conversational base on [chatml format](https://github.com/MicrosoftDocs/azure-docs/blob/main/articles/ai-services/openai/includes/chat-markup-language.md#working-with-chat-markup-language-chatml) we identified the special tokens so the model could understand the different roles in the conversation Example ``` <bos><\|im_start\|>system You are Gemma.<\|im_end\|> <\|im_start\|>user Hello, how are you?<\|im_end\|> <\|im_start\|>assistant I'm doing great. How can I help you today?<\|im_end\|>\n<eos> ``` So we used [Phil Schmid's gemma chatml tokenizer](https://huggingface.co/philschmid/gemma-tokenizer-chatml) to adapt our dataset for training #### Training Hyperparameters <!-- Enumerar los valores de los hiperparámetros de entrenamiento. --> The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 66 - gradient_accumulation_steps: 2 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 3 - Training regime: <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results <!-- Enlazar aquí los scripts/notebooks de evaluación y especificar los resultados. --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly. --> <!-- Rellenar la información de la lista y calcular las emisiones con la página mencionada. --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: 1 X RTX4090 - Hours used: 4 - Cloud Provider: Vast.ai - Compute Region: West Europe - Carbon Emitted: 350W x 4h = 1.4 kWh x 0.57 kg eq. CO2/kWh = 0.8 kg eq. CO2 ## Technical Specifications <!-- Esta sección es opcional porque seguramente ya habéis mencionado estos detalles más arriba, igualmente está bien incluirlos aquí de nuevo como bullet points a modo de resumen. --> ### Model Architecture and Objective The base model is [google/gemma-7b](https://huggingface.co/google/gemma-7b) finetuned in 4-bit. ### Compute Infrastructure #### Hardware <!-- Indicar el hardware utilizado, podéis agradecer aquí a quien lo patrocinó. --> 1 x RTX4090 GPU by Vast.ai. #### Software <!-- Enumerar las librerías utilizadas (e.g. transformers, distilabel). --> Libraries: - transformers - bitsandbytes - accelerate - xformers - trl - peft - wandb ## License <!-- Indicar bajo qué licencia se libera el modelo explicando, si no es apache 2.0, a qué se debe la licencia más restrictiva (i.e. herencia de las licencias del modelo pre-entrenado o de los datos utilizados). --> This model is under the license of the Gemma models by Google. Link to consent: https://www.kaggle.com/models/google/gemma/license/consent ## Citation BibTeX: [More Information Needed] <!-- Aquí tenéis un ejemplo de cita de un dataset que podéis adaptar: ``` @software{benallal2024cosmopedia, author = {Ben Allal, Loubna and Lozhkov, Anton and Penedo, Guilherme and Wolf, Thomas and von Werra, Leandro}, title = {Cosmopedia}, month = February, year = 2024, url = {https://huggingface.co/datasets/HuggingFaceTB/cosmopedia} } ``` - benallal2024cosmopedia -> nombre + año + nombre del modelo - author: lista de miembros del equipo - title: nombre del modelo - year: año - url: enlace al modelo --> ``` @software{somosnlp2024asistenciarefugiados, author = {Alvaro Hidalgo, Eduardo Muñoz, Teresa Martín}, title = {gemma-7b-it-legal-refugiados-es}, month = April, year = 2024, url = {somosnlp/gemma-7b-it-legal-refugee-v0.1.1} } ``` ## More Information <!-- Indicar aquí que el marco en el que se desarrolló el proyecto, en esta sección podéis incluir agradecimientos y más información sobre los miembros del equipo. Podéis adaptar el ejemplo a vuestro gusto. --> This project was developed during the [Hackathon #Somos600M](https://somosnlp.org/hackathon) organized by SomosNLP. The model was trained using GPUs sponsored by HuggingFace. Team: [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim) <!-- - [Name 1](Link to Hugging Face profile) - [Name 2](Link to Hugging Face profile) --> ## Contact [optional] <!-- Email de contacto para´posibles preguntas sobre el modelo. -->	{"base_model": "google/gemma-7b", "datasets": ["somosnlp/instruct-legal-refugiados-es"], "language": ["es"], "library_name": "transformers, pe", "license": "apache-2.0", "tags": ["trl", "sft", "generated_from_trainer"]}	task	[ "QUESTION_ANSWERING" ]	44,597
MugheesAwan11/bge-base-securiti-dataset-1-v10	MugheesAwan11	sentence-similarity	[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:900", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss", "en", "arxiv:1908.10084", "arxiv:2205.13147", "arxiv:1705.00652", "base_model:BAAI/bge-base-en-v1.5", "base_model:finetune:BAAI/bge-base-en-v1.5", "license:apache-2.0", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	2024-06-13T17:25:39Z	2024-06-13T17:25:56+00:00	6	0	--- base_model: BAAI/bge-base-en-v1.5 datasets: [] language: - en library_name: sentence-transformers license: apache-2.0 metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:900 - loss:MatryoshkaLoss - loss:MultipleNegativesRankingLoss widget: - source_sentence: Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about Securiti, our mission and history Partner Program Join our Partner Program Contact Us Contact us to learn more or schedule a demo News Coverage Read about Securiti in the news Press Releases Find our latest press releases Careers Join the sentences: - What is the purpose of tracking changes and transformations of data throughout its lifecycle? - What is the role of ePD in the European privacy regime and its relation to GDPR? - How can data governance be optimized using granular insights? - source_sentence: Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud \| Data Mapping \| DSR Automation \| Assessment Automation \| Vendor Assessment \| Breach Management \| Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data \| People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent \| Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator , View Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud \| Data Mapping \| DSR Automation \| Assessment Automation \| Vendor Assessment \| Breach Management \| Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data \| People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent \| Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls sentences: - What is the purpose of Asset and Data Discovery in data governance and security? - Which EU member states have strict cyber laws? - What is the obligation for organizations to provide Data Protection Impact Assessments (DPIAs) under the LGPD? - source_sentence: 'which the data is processed. Right to Access: Data subjects have the right to obtain confirmation whether or not the controller holds personal data about them, access their personal data, and obtain descriptions of data recipients. Right to Rectification : Under the right to rectification, data subjects can request the correction of their data. Right to Erasure: Data subjects have the right to request the erasure and destruction of the data that is no longer needed by the organization. Right to Object: The data subject has the right to prevent the data controller from processing personal data if such processing causes or is likely to cause unwarranted damage or distress to the data subject. Right not to be Subjected to Automated Decision-Making : The data subject has the right to not be subject to automated decision-making that significantly affects the individual. ## Facts related to Ghana’s Data Protection Act 2012 1 While processing personal data, organizations must comply with eight privacy principles: lawfulness of processing, data quality, security measures, accountability, purpose specification, purpose limitation, openness, and data subject participation. 2 In the event of a security breach, the data controller shall take measures to prevent the breach and notify the Commission and the data subject about the breach as soon as reasonably practicable after the discovery of the breach. 3 The DPA specifies lawful grounds for data processing, including data subject’s consent, the performance of a contract, the interest of data subject and public interest, lawful obligations, and the legitimate interest of the data controller. 4 The DPA requires data controllers to register with the Data Protection Commission (DPC). 5 The DPA provides varying fines and terms of imprisonment according to the severity and sensitivity of the violation, such as any person who sells personal data may get fined up to 2500 penalty units or up to five years imprisonment or both. ### Forrester Names Securiti a Leader in the Privacy Management Wave Q4, 2021 Read the Report ### Securiti named a Leader in the IDC MarketScape for Data Privacy Compliance Software Read the Report At Securiti, our mission is to enable enterprises to safely harness the incredible power of data and the cloud by controlling the complex security, privacy and compliance risks. Copyright (C) 2023 Securiti Sitem' sentences: - What information is required for data subjects regarding data transfers under the GDPR, including personal data categories, data recipients, retention period, and automated decision making? - What privacy principles must organizations follow when processing personal data under Ghana's Data Protection Act 2012? - What is the purpose of Thailand's PDPA? - source_sentence: 'consumer has the right to have his/her personal data stored or processed by the data controller be deleted. ## Portability The consumer has a right to obtain a copy of his/her personal data in a portable, technically feasible and readily usable format that allows the consumer to transmit the data to another controller without hindrance. ## Opt out The consumer has the right to opt out of the processing of the personal data for purposes of targeted advertising, the sale of personal data, or profiling in furtherance of decisions that produce legal or similarly significant effects concerning the consumer. Time period to fulfill DSR request: All data subject rights’ requests (DSR requests) must be fulfilled by the data controller within a 45 day period. Extension in time period: data controllers may seek for an extension of 45 days in fulfilling the request depending on the complexity and number of the consumer''s requests. Denial of DSR request: If a DSR request is to be denied, the data controller must inform the consumer of the reasons within a 45 days period. Appeal against refusal: Consumers have a right to appeal the decision for refusal of grant of the DSR request. The appeal must be decided within 45 days but the time period can be further extended by 60 additional days. Limitation of DSR requests per year: Requests for data portability may be made only twice in a year. Charges: DSR requests must be fulfilled free of charge once in a year. Any subsequent request within a 12 month period can be charged. Authentication: A data controller is not to respond to a consumer request unless it can authenticate the request using reasonably commercial means. A data controller can request additional information from the consumer for the purposes of authenticating the request. ## Who must comply? CPA applies to all data controllers who conduct business in Colorado or produce or deliver commercial products or services that are intentionally targeted to residents of Colorado if they match any one or both of these conditions: If they control or process the personal data of 100,000 consumers or more during a calendar year; or If they derive revenue or receive a discount on the price of goods or services from the sale of personal data and process or control the personal data of 25,000' sentences: - What is the US California CCPA and how does it relate to data privacy regulations? - What does the People Data Graph serve in terms of privacy, security, and governance? - What rights does a consumer have regarding the portability of their personal data? - source_sentence: 'PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law''s provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany''s Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR''s €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country''s data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company''s employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users'' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory' sentences: - What are the benefits of automating compliance with global privacy regulations for data protection and control? - What is required for an official complaint to be filed under Germany's Federal Data Protection Act? - Why is tracking data lineage important for data management and security? model-index: - name: SentenceTransformer based on BAAI/bge-base-en-v1.5 results: - task: type: information-retrieval name: Information Retrieval dataset: name: dim 512 type: dim_512 metrics: - type: cosine_accuracy@1 value: 0.08 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.29 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.48 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.65 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.08 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.09666666666666668 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.09599999999999997 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.06499999999999999 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.08 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.29 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.48 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.65 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.3356834483699582 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.23805952380952378 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.25373588653956675 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 256 type: dim_256 metrics: - type: cosine_accuracy@1 value: 0.09 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.33 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.52 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.68 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.09 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.11 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.10399999999999998 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.06799999999999998 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.09 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.33 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.52 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.68 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.35403179411423247 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.2524960317460317 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.26470102220887337 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 128 type: dim_128 metrics: - type: cosine_accuracy@1 value: 0.09 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.27 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.45 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.65 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.09 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.09 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.09 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.06499999999999999 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.09 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.27 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.45 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.65 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.33203261209382817 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.23417063492063486 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.24858408269645846 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 64 type: dim_64 metrics: - type: cosine_accuracy@1 value: 0.06 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.23 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.44 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.57 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.06 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.07666666666666666 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.08799999999999997 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.056999999999999995 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.06 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.23 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.44 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.57 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.28544770610641695 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.19726587301587298 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.21493811628701745 name: Cosine Map@100 --- # SentenceTransformer based on BAAI/bge-base-en-v1.5 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5) <!-- at revision a5beb1e3e68b9ab74eb54cfd186867f64f240e1a --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 768 tokens - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> - Language: en - License: apache-2.0 ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("MugheesAwan11/bge-base-securiti-dataset-1-v10") # Run inference sentences = [ "PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law's provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany's Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR's €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country's data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company's employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory", "What is required for an official complaint to be filed under Germany's Federal Data Protection Act?", 'Why is tracking data lineage important for data management and security?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_512` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.08 \| \| cosine_accuracy@3 \| 0.29 \| \| cosine_accuracy@5 \| 0.48 \| \| cosine_accuracy@10 \| 0.65 \| \| cosine_precision@1 \| 0.08 \| \| cosine_precision@3 \| 0.0967 \| \| cosine_precision@5 \| 0.096 \| \| cosine_precision@10 \| 0.065 \| \| cosine_recall@1 \| 0.08 \| \| cosine_recall@3 \| 0.29 \| \| cosine_recall@5 \| 0.48 \| \| cosine_recall@10 \| 0.65 \| \| cosine_ndcg@10 \| 0.3357 \| \| cosine_mrr@10 \| 0.2381 \| \| cosine_map@100 \| 0.2537 \| #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.09 \| \| cosine_accuracy@3 \| 0.33 \| \| cosine_accuracy@5 \| 0.52 \| \| cosine_accuracy@10 \| 0.68 \| \| cosine_precision@1 \| 0.09 \| \| cosine_precision@3 \| 0.11 \| \| cosine_precision@5 \| 0.104 \| \| cosine_precision@10 \| 0.068 \| \| cosine_recall@1 \| 0.09 \| \| cosine_recall@3 \| 0.33 \| \| cosine_recall@5 \| 0.52 \| \| cosine_recall@10 \| 0.68 \| \| cosine_ndcg@10 \| 0.354 \| \| cosine_mrr@10 \| 0.2525 \| \| cosine_map@100 \| 0.2647 \| #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.09 \| \| cosine_accuracy@3 \| 0.27 \| \| cosine_accuracy@5 \| 0.45 \| \| cosine_accuracy@10 \| 0.65 \| \| cosine_precision@1 \| 0.09 \| \| cosine_precision@3 \| 0.09 \| \| cosine_precision@5 \| 0.09 \| \| cosine_precision@10 \| 0.065 \| \| cosine_recall@1 \| 0.09 \| \| cosine_recall@3 \| 0.27 \| \| cosine_recall@5 \| 0.45 \| \| cosine_recall@10 \| 0.65 \| \| cosine_ndcg@10 \| 0.332 \| \| cosine_mrr@10 \| 0.2342 \| \| cosine_map@100 \| 0.2486 \| #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.06 \| \| cosine_accuracy@3 \| 0.23 \| \| cosine_accuracy@5 \| 0.44 \| \| cosine_accuracy@10 \| 0.57 \| \| cosine_precision@1 \| 0.06 \| \| cosine_precision@3 \| 0.0767 \| \| cosine_precision@5 \| 0.088 \| \| cosine_precision@10 \| 0.057 \| \| cosine_recall@1 \| 0.06 \| \| cosine_recall@3 \| 0.23 \| \| cosine_recall@5 \| 0.44 \| \| cosine_recall@10 \| 0.57 \| \| cosine_ndcg@10 \| 0.2854 \| \| cosine_mrr@10 \| 0.1973 \| \| cosine_map@100 \| 0.2149 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 900 training samples * Columns: <code>positive</code> and <code>anchor</code> * Approximate statistics based on the first 1000 samples: \| \| positive \| anchor \| \|:--------\|:--------------------------------------------------------------------------------------\|:----------------------------------------------------------------------------------\| \| type \| string \| string \| \| details \| <ul><li>min: 159 tokens</li><li>mean: 445.26 tokens</li><li>max: 512 tokens</li></ul> \| <ul><li>min: 7 tokens</li><li>mean: 22.05 tokens</li><li>max: 82 tokens</li></ul> \| * Samples: \| positive \| anchor \| \|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\|:--------------------------------------------------------------------------------------------------------\| \| <code>orra The Andorra personal data protection act came into force on May 17, 2022, by the Andorra Data Protection Authority (ADPA). Learn more about Andorra PDPA ### United Kingdom The UK Data Protection Act (DPA) 2018 is the amended version of the Data Protection Act that was passed in 1998. The DPA 2018 implements the GDPR with several additions and restrictions. Learn more about UK DPA ### Botswana The Botswana Data Protection came into effect on October 15, 2021 after the issuance of the Data Protection Act (Commencement Date) Order 2021 by the Minister of Presidential Affairs, Governance and Public Administration. Learn more about Botswana DPA ### Zambia On March 31, 2021, the Zambian parliament formally passed the Data Protection Act No. 3 of 2021 and the Electronic Communications and Transactions Act No. 4 of 2021. Learn more about Zambia DPA ### Jamaica On November 30, 2020, the First Schedule of the Data Protection Act No. 7 of 2020 came into effect following the publication of Supplement No. 160 of Volume CXLIV in the Jamaica Gazette Supplement. Learn more about Jamaica DPA ### Belarus The Law on Personal Data Protection of May 7, 2021, No. 99-Z, entered into effect within Belarus on November 15, 2021. Learn more about Belarus DPA ### Russian Federation The primary Russian law on data protection, Federal Law No. 152-FZ has been in effect since July 2006. Learn more ### Eswatini On March 4, 2022, the Eswatini Communications Commission published the Data Protection Act No. 5 of 2022, simultaneously announcing its immediate enforcement. Learn more ### Oman The Royal Decree 6/2022 promulgating the Personal Data Protection Law (PDPL) was passed on February 9, 2022. Learn more ### Sri Lanka Sri Lanka's parliament formally passed the Personal Data Protection Act (PDPA), No. 9 Of 2022, on March 19, 2022. Learn more ### Kuwait Kuwait's DPPR was formally introduced by the CITRA to ensure the Gulf country's data privacy infrastructure. Learn more ### Brunei Darussalam The draft Personal Data Protection Order is Brunei’s primary data protection law which came into effect in 2022. Learn more ### India India’</code> \| <code>What is the name of India's data protection law before May 17, 2022?</code> \| \| <code>the affected data subjects and regulatory authority about the breach and whether any of their information has been compromised as a result. ### Data Protection Impact Assessment There is no requirement for conducting data protection impact assessment under the PDPA. ### Record of Processing Activities A data controller must keep and maintain a record of any privacy notice, data subject request, or any other information relating to personal data processed by him in the form and manner that may be determined by the regulatory authority. ### Cross Border Data Transfer Requirements The PDPA provides that personal data can be transferred out of Malaysia only when the recipient country is specified as adequate in the Official Gazette. The personal data of data subjects can not be disclosed without the consent of the data subject. The PDPA provides the following exceptions to the cross border data transfer requirements: Where the consent of data subject is obtained for transfer; or Where the transfer is necessary for the performance of contract between the parties; The transfer is for the purpose of any legal proceedings or for the purpose of obtaining legal advice or for establishing, exercising or defending legal rights; The data user has taken all reasonable precautions and exercised all due diligence to ensure that the personal data will not in that place be processed in any manner which, if that place is Malaysia, would be a contravention of this PDPA; The transfer is necessary in order to protect the vital interests of the data subject; or The transfer is necessary as being in the public interest in circumstances as determined by the Minister. ## Data Subject Rights The data subjects or the person whose data is being collected has certain rights under the PDPA. The most prominent rights can be categorized under the following: ## Right to withdraw consent The PDPA, like some of the other landmark data protection laws such as CPRA and GDPR gives data subjects the right to revoke their consent at any time by way of written notice from having their data collected processed. ## Right to access and rectification As per this right, anyone whose data has been collected has the right to request to review their personal data and have it updated. The onus is on the data handlers to respond to such a request as soon as possible while also making it easier for data subjects on how they can request access to their personal data. ## Right to data portability Data subjects have the right to request that their data be stored in a manner where it</code> \| <code>What is the requirement for conducting a data protection impact assessment under the PDPA?</code> \| \| <code>more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about</code> \| <code>What is Data Subject Request Automation?</code> \| * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 5 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 5 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs \| Epoch \| Step \| Training Loss \| dim_128_cosine_map@100 \| dim_256_cosine_map@100 \| dim_512_cosine_map@100 \| dim_64_cosine_map@100 \| \|:-------:\|:-------:\|:-------------:\|:----------------------:\|:----------------------:\|:----------------------:\|:---------------------:\| \| 0.3448 \| 10 \| 7.4297 \| - \| - \| - \| - \| \| 0.6897 \| 20 \| 5.5127 \| - \| - \| - \| - \| \| 1.0 \| 29 \| - \| 0.2399 \| 0.2435 \| 0.2579 \| 0.1837 \| \| 1.0345 \| 30 \| 4.8788 \| - \| - \| - \| - \| \| 1.3793 \| 40 \| 4.0614 \| - \| - \| - \| - \| \| 1.7241 \| 50 \| 3.3471 \| - \| - \| - \| - \| \| 2.0 \| 58 \| - \| 0.2373 \| 0.2510 \| 0.2545 \| 0.1964 \| \| 2.0690 \| 60 \| 3.104 \| - \| - \| - \| - \| \| 2.4138 \| 70 \| 2.695 \| - \| - \| - \| - \| \| 2.7586 \| 80 \| 2.2038 \| - \| - \| - \| - \| \| 3.0 \| 87 \| - \| 0.2416 \| 0.2630 \| 0.2587 \| 0.2121 \| \| 3.1034 \| 90 \| 2.2576 \| - \| - \| - \| - \| \| 3.4483 \| 100 \| 2.1552 \| - \| - \| - \| - \| \| 3.7931 \| 110 \| 1.8199 \| - \| - \| - \| - \| \| 4.0 \| 116 \| - \| 0.2429 \| 0.2613 \| 0.2546 \| 0.2098 \| \| 4.1379 \| 120 \| 1.9192 \| - \| - \| - \| - \| \| 4.4828 \| 130 \| 1.7221 \| - \| - \| - \| - \| \| 4.8276 \| 140 \| 1.6878 \| - \| - \| - \| - \| \| 5.0 \| 145 \| - \| 0.2486 \| 0.2647 \| 0.2537 \| 0.2149 \| * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.1.2+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# SentenceTransformer based on BAAI/bge-base-en-v1.5 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - Model Type: Sentence Transformer - Base model: [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5) <!-- at revision a5beb1e3e68b9ab74eb54cfd186867f64f240e1a --> - Maximum Sequence Length: 512 tokens - Output Dimensionality: 768 tokens - Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> - Language: en - License: apache-2.0 ### Model Sources - Documentation: [Sentence Transformers Documentation](https://sbert.net) - Repository: [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("MugheesAwan11/bge-base-securiti-dataset-1-v10") # Run inference sentences = [ "PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law's provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany's Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR's €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country's data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company's employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory", "What is required for an official complaint to be filed under Germany's Federal Data Protection Act?", 'Why is tracking data lineage important for data management and security?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` <!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_512` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.08 \| \| cosine_accuracy@3 \| 0.29 \| \| cosine_accuracy@5 \| 0.48 \| \| cosine_accuracy@10 \| 0.65 \| \| cosine_precision@1 \| 0.08 \| \| cosine_precision@3 \| 0.0967 \| \| cosine_precision@5 \| 0.096 \| \| cosine_precision@10 \| 0.065 \| \| cosine_recall@1 \| 0.08 \| \| cosine_recall@3 \| 0.29 \| \| cosine_recall@5 \| 0.48 \| \| cosine_recall@10 \| 0.65 \| \| cosine_ndcg@10 \| 0.3357 \| \| cosine_mrr@10 \| 0.2381 \| \| cosine_map@100 \| 0.2537 \| #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.09 \| \| cosine_accuracy@3 \| 0.33 \| \| cosine_accuracy@5 \| 0.52 \| \| cosine_accuracy@10 \| 0.68 \| \| cosine_precision@1 \| 0.09 \| \| cosine_precision@3 \| 0.11 \| \| cosine_precision@5 \| 0.104 \| \| cosine_precision@10 \| 0.068 \| \| cosine_recall@1 \| 0.09 \| \| cosine_recall@3 \| 0.33 \| \| cosine_recall@5 \| 0.52 \| \| cosine_recall@10 \| 0.68 \| \| cosine_ndcg@10 \| 0.354 \| \| cosine_mrr@10 \| 0.2525 \| \| cosine_map@100 \| 0.2647 \| #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.09 \| \| cosine_accuracy@3 \| 0.27 \| \| cosine_accuracy@5 \| 0.45 \| \| cosine_accuracy@10 \| 0.65 \| \| cosine_precision@1 \| 0.09 \| \| cosine_precision@3 \| 0.09 \| \| cosine_precision@5 \| 0.09 \| \| cosine_precision@10 \| 0.065 \| \| cosine_recall@1 \| 0.09 \| \| cosine_recall@3 \| 0.27 \| \| cosine_recall@5 \| 0.45 \| \| cosine_recall@10 \| 0.65 \| \| cosine_ndcg@10 \| 0.332 \| \| cosine_mrr@10 \| 0.2342 \| \| cosine_map@100 \| 0.2486 \| #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) \| Metric \| Value \| \|:--------------------\|:-----------\| \| cosine_accuracy@1 \| 0.06 \| \| cosine_accuracy@3 \| 0.23 \| \| cosine_accuracy@5 \| 0.44 \| \| cosine_accuracy@10 \| 0.57 \| \| cosine_precision@1 \| 0.06 \| \| cosine_precision@3 \| 0.0767 \| \| cosine_precision@5 \| 0.088 \| \| cosine_precision@10 \| 0.057 \| \| cosine_recall@1 \| 0.06 \| \| cosine_recall@3 \| 0.23 \| \| cosine_recall@5 \| 0.44 \| \| cosine_recall@10 \| 0.57 \| \| cosine_ndcg@10 \| 0.2854 \| \| cosine_mrr@10 \| 0.1973 \| \| cosine_map@100 \| 0.2149 \| <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 900 training samples * Columns: <code>positive</code> and <code>anchor</code> * Approximate statistics based on the first 1000 samples: \| \| positive \| anchor \| \|:--------\|:--------------------------------------------------------------------------------------\|:----------------------------------------------------------------------------------\| \| type \| string \| string \| \| details \| <ul><li>min: 159 tokens</li><li>mean: 445.26 tokens</li><li>max: 512 tokens</li></ul> \| <ul><li>min: 7 tokens</li><li>mean: 22.05 tokens</li><li>max: 82 tokens</li></ul> \| * Samples: \| positive \| anchor \| \|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\|:--------------------------------------------------------------------------------------------------------\| \| <code>orra The Andorra personal data protection act came into force on May 17, 2022, by the Andorra Data Protection Authority (ADPA). Learn more about Andorra PDPA ### United Kingdom The UK Data Protection Act (DPA) 2018 is the amended version of the Data Protection Act that was passed in 1998. The DPA 2018 implements the GDPR with several additions and restrictions. Learn more about UK DPA ### Botswana The Botswana Data Protection came into effect on October 15, 2021 after the issuance of the Data Protection Act (Commencement Date) Order 2021 by the Minister of Presidential Affairs, Governance and Public Administration. Learn more about Botswana DPA ### Zambia On March 31, 2021, the Zambian parliament formally passed the Data Protection Act No. 3 of 2021 and the Electronic Communications and Transactions Act No. 4 of 2021. Learn more about Zambia DPA ### Jamaica On November 30, 2020, the First Schedule of the Data Protection Act No. 7 of 2020 came into effect following the publication of Supplement No. 160 of Volume CXLIV in the Jamaica Gazette Supplement. Learn more about Jamaica DPA ### Belarus The Law on Personal Data Protection of May 7, 2021, No. 99-Z, entered into effect within Belarus on November 15, 2021. Learn more about Belarus DPA ### Russian Federation The primary Russian law on data protection, Federal Law No. 152-FZ has been in effect since July 2006. Learn more ### Eswatini On March 4, 2022, the Eswatini Communications Commission published the Data Protection Act No. 5 of 2022, simultaneously announcing its immediate enforcement. Learn more ### Oman The Royal Decree 6/2022 promulgating the Personal Data Protection Law (PDPL) was passed on February 9, 2022. Learn more ### Sri Lanka Sri Lanka's parliament formally passed the Personal Data Protection Act (PDPA), No. 9 Of 2022, on March 19, 2022. Learn more ### Kuwait Kuwait's DPPR was formally introduced by the CITRA to ensure the Gulf country's data privacy infrastructure. Learn more ### Brunei Darussalam The draft Personal Data Protection Order is Brunei’s primary data protection law which came into effect in 2022. Learn more ### India India’</code> \| <code>What is the name of India's data protection law before May 17, 2022?</code> \| \| <code>the affected data subjects and regulatory authority about the breach and whether any of their information has been compromised as a result. ### Data Protection Impact Assessment There is no requirement for conducting data protection impact assessment under the PDPA. ### Record of Processing Activities A data controller must keep and maintain a record of any privacy notice, data subject request, or any other information relating to personal data processed by him in the form and manner that may be determined by the regulatory authority. ### Cross Border Data Transfer Requirements The PDPA provides that personal data can be transferred out of Malaysia only when the recipient country is specified as adequate in the Official Gazette. The personal data of data subjects can not be disclosed without the consent of the data subject. The PDPA provides the following exceptions to the cross border data transfer requirements: Where the consent of data subject is obtained for transfer; or Where the transfer is necessary for the performance of contract between the parties; The transfer is for the purpose of any legal proceedings or for the purpose of obtaining legal advice or for establishing, exercising or defending legal rights; The data user has taken all reasonable precautions and exercised all due diligence to ensure that the personal data will not in that place be processed in any manner which, if that place is Malaysia, would be a contravention of this PDPA; The transfer is necessary in order to protect the vital interests of the data subject; or The transfer is necessary as being in the public interest in circumstances as determined by the Minister. ## Data Subject Rights The data subjects or the person whose data is being collected has certain rights under the PDPA. The most prominent rights can be categorized under the following: ## Right to withdraw consent The PDPA, like some of the other landmark data protection laws such as CPRA and GDPR gives data subjects the right to revoke their consent at any time by way of written notice from having their data collected processed. ## Right to access and rectification As per this right, anyone whose data has been collected has the right to request to review their personal data and have it updated. The onus is on the data handlers to respond to such a request as soon as possible while also making it easier for data subjects on how they can request access to their personal data. ## Right to data portability Data subjects have the right to request that their data be stored in a manner where it</code> \| <code>What is the requirement for conducting a data protection impact assessment under the PDPA?</code> \| \| <code>more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about</code> \| <code>What is Data Subject Request Automation?</code> \| * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 5 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 5 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs \| Epoch \| Step \| Training Loss \| dim_128_cosine_map@100 \| dim_256_cosine_map@100 \| dim_512_cosine_map@100 \| dim_64_cosine_map@100 \| \|:-------:\|:-------:\|:-------------:\|:----------------------:\|:----------------------:\|:----------------------:\|:---------------------:\| \| 0.3448 \| 10 \| 7.4297 \| - \| - \| - \| - \| \| 0.6897 \| 20 \| 5.5127 \| - \| - \| - \| - \| \| 1.0 \| 29 \| - \| 0.2399 \| 0.2435 \| 0.2579 \| 0.1837 \| \| 1.0345 \| 30 \| 4.8788 \| - \| - \| - \| - \| \| 1.3793 \| 40 \| 4.0614 \| - \| - \| - \| - \| \| 1.7241 \| 50 \| 3.3471 \| - \| - \| - \| - \| \| 2.0 \| 58 \| - \| 0.2373 \| 0.2510 \| 0.2545 \| 0.1964 \| \| 2.0690 \| 60 \| 3.104 \| - \| - \| - \| - \| \| 2.4138 \| 70 \| 2.695 \| - \| - \| - \| - \| \| 2.7586 \| 80 \| 2.2038 \| - \| - \| - \| - \| \| 3.0 \| 87 \| - \| 0.2416 \| 0.2630 \| 0.2587 \| 0.2121 \| \| 3.1034 \| 90 \| 2.2576 \| - \| - \| - \| - \| \| 3.4483 \| 100 \| 2.1552 \| - \| - \| - \| - \| \| 3.7931 \| 110 \| 1.8199 \| - \| - \| - \| - \| \| 4.0 \| 116 \| - \| 0.2429 \| 0.2613 \| 0.2546 \| 0.2098 \| \| 4.1379 \| 120 \| 1.9192 \| - \| - \| - \| - \| \| 4.4828 \| 130 \| 1.7221 \| - \| - \| - \| - \| \| 4.8276 \| 140 \| 1.6878 \| - \| - \| - \| - \| \| 5.0 \| 145 \| - \| 0.2486 \| 0.2647 \| 0.2537 \| 0.2149 \| * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.1.2+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "BAAI/bge-base-en-v1.5", "datasets": [], "language": ["en"], "library_name": "sentence-transformers", "license": "apache-2.0", "metrics": ["cosine_accuracy@1", "cosine_accuracy@3", "cosine_accuracy@5", "cosine_accuracy@10", "cosine_precision@1", "cosine_precision@3", "cosine_precision@5", "cosine_precision@10", "cosine_recall@1", "cosine_recall@3", "cosine_recall@5", "cosine_recall@10", "cosine_ndcg@10", "cosine_mrr@10", "cosine_map@100"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:900", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss"], "widget": [{"source_sentence": "Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \\+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about Securiti, our mission and history Partner Program Join our Partner Program Contact Us Contact us to learn more or schedule a demo News Coverage Read about Securiti in the news Press Releases Find our latest press releases Careers Join the", "sentences": ["What is the purpose of tracking changes and transformations of data throughout its lifecycle?", "What is the role of ePD in the European privacy regime and its relation to GDPR?", "How can data governance be optimized using granular insights?"]}, {"source_sentence": "Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud \| Data Mapping \| DSR Automation \| Assessment Automation \| Vendor Assessment \| Breach Management \| Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data \| People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent \| Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator , View Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud \| Data Mapping \| DSR Automation \| Assessment Automation \| Vendor Assessment \| Breach Management \| Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data \| People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent \| Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls", "sentences": ["What is the purpose of Asset and Data Discovery in data governance and security?", "Which EU member states have strict cyber laws?", "What is the obligation for organizations to provide Data Protection Impact Assessments (DPIAs) under the LGPD?"]}, {"source_sentence": "which the data is processed. Right to Access: Data subjects have the right to obtain confirmation whether or not the controller holds personal data about them, access their personal data, and obtain descriptions of data recipients. Right to Rectification : Under the right to rectification, data subjects can request the correction of their data. Right to Erasure: Data subjects have the right to request the erasure and destruction of the data that is no longer needed by the organization. Right to Object: The data subject has the right to prevent the data controller from processing personal data if such processing causes or is likely to cause unwarranted damage or distress to the data subject. Right not to be Subjected to Automated Decision-Making : The data subject has the right to not be subject to automated decision-making that significantly affects the individual. ## Facts related to Ghana’s Data Protection Act 2012 1 While processing personal data, organizations must comply with eight privacy principles: lawfulness of processing, data quality, security measures, accountability, purpose specification, purpose limitation, openness, and data subject participation. 2 In the event of a security breach, the data controller shall take measures to prevent the breach and notify the Commission and the data subject about the breach as soon as reasonably practicable after the discovery of the breach. 3 The DPA specifies lawful grounds for data processing, including data subject’s consent, the performance of a contract, the interest of data subject and public interest, lawful obligations, and the legitimate interest of the data controller. 4 The DPA requires data controllers to register with the Data Protection Commission (DPC). 5 The DPA provides varying fines and terms of imprisonment according to the severity and sensitivity of the violation, such as any person who sells personal data may get fined up to 2500 penalty units or up to five years imprisonment or both. ### Forrester Names Securiti a Leader in the Privacy Management Wave Q4, 2021 Read the Report ### Securiti named a Leader in the IDC MarketScape for Data Privacy Compliance Software Read the Report At Securiti, our mission is to enable enterprises to safely harness the incredible power of data and the cloud by controlling the complex security, privacy and compliance risks. Copyright (C) 2023 Securiti Sitem", "sentences": ["What information is required for data subjects regarding data transfers under the GDPR, including personal data categories, data recipients, retention period, and automated decision making?", "What privacy principles must organizations follow when processing personal data under Ghana's Data Protection Act 2012?", "What is the purpose of Thailand's PDPA?"]}, {"source_sentence": "consumer has the right to have his/her personal data stored or processed by the data controller be deleted. ## Portability The consumer has a right to obtain a copy of his/her personal data in a portable, technically feasible and readily usable format that allows the consumer to transmit the data to another controller without hindrance. ## Opt out The consumer has the right to opt out of the processing of the personal data for purposes of targeted advertising, the sale of personal data, or profiling in furtherance of decisions that produce legal or similarly significant effects concerning the consumer. Time period to fulfill DSR request: All data subject rights’ requests (DSR requests) must be fulfilled by the data controller within a 45 day period. Extension in time period: data controllers may seek for an extension of 45 days in fulfilling the request depending on the complexity and number of the consumer's requests. Denial of DSR request: If a DSR request is to be denied, the data controller must inform the consumer of the reasons within a 45 days period. Appeal against refusal: Consumers have a right to appeal the decision for refusal of grant of the DSR request. The appeal must be decided within 45 days but the time period can be further extended by 60 additional days. Limitation of DSR requests per year: Requests for data portability may be made only twice in a year. Charges: DSR requests must be fulfilled free of charge once in a year. Any subsequent request within a 12 month period can be charged. Authentication: A data controller is not to respond to a consumer request unless it can authenticate the request using reasonably commercial means. A data controller can request additional information from the consumer for the purposes of authenticating the request. ## Who must comply? CPA applies to all data controllers who conduct business in Colorado or produce or deliver commercial products or services that are intentionally targeted to residents of Colorado if they match any one or both of these conditions: If they control or process the personal data of 100,000 consumers or more during a calendar year; or If they derive revenue or receive a discount on the price of goods or services from the sale of personal data and process or control the personal data of 25,000", "sentences": ["What is the US California CCPA and how does it relate to data privacy regulations?", "What does the People Data Graph serve in terms of privacy, security, and governance?", "What rights does a consumer have regarding the portability of their personal data?"]}, {"source_sentence": "PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law's provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany's Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR's €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country's data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company's employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory", "sentences": ["What are the benefits of automating compliance with global privacy regulations for data protection and control?", "What is required for an official complaint to be filed under Germany's Federal Data Protection Act?", "Why is tracking data lineage important for data management and security?"]}], "model-index": [{"name": "SentenceTransformer based on BAAI/bge-base-en-v1.5", "results": [{"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 512", "type": "dim_512"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.08, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.29, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.48, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.65, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.08, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.09666666666666668, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.09599999999999997, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.06499999999999999, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.08, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.29, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.48, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.65, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.3356834483699582, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.23805952380952378, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.25373588653956675, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 256", "type": "dim_256"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.09, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.33, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.52, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.68, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.09, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.11, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.10399999999999998, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.06799999999999998, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.09, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.33, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.52, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.68, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.35403179411423247, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.2524960317460317, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.26470102220887337, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 128", "type": "dim_128"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.09, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.27, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.45, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.65, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.09, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.09, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.09, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.06499999999999999, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.09, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.27, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.45, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.65, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.33203261209382817, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.23417063492063486, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.24858408269645846, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 64", "type": "dim_64"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.06, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.23, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.44, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.57, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.06, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.07666666666666666, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.08799999999999997, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.056999999999999995, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.06, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.23, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.44, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.57, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.28544770610641695, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.19726587301587298, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.21493811628701745, "name": "Cosine Map@100"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,598
danielpark/ko-llama-2-jindo-13b-instruct	danielpark	text-generation	[ "peft", "pytorch", "llama", "dsdanielpark", "llama2", "instruct", "instruction", "jindo", "korean", "translation", "13b", "text-generation", "en", "ko", "dataset:korean-jindo-dataset.json", "arxiv:2307.09288", "region:us" ]	2023-07-27T07:10:35Z	2023-11-13T08:29:33+00:00	0	2	--- datasets: - korean-jindo-dataset.json language: - en - ko library_name: peft pipeline_tag: text-generation tags: - dsdanielpark - llama2 - instruct - instruction - jindo - korean - translation - 13b --- # Since this model is still under development, I recommend not using it until it reaches the development stage 5. Development Status :: 2 - Pre-Alpha <br> Developed by MinWoo Park, 2023, Seoul, South Korea. [Contact: [email protected]](mailto:[email protected]). [![Hits](https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fhuggingface.co%2Fdanielpark%2Fko-llama-2-jindo-13b-instruct&count_bg=%23000000&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=views&edge_flat=false)](https://hits.seeyoufarm.com) # danielpark/llama2-jindo-13b-instruct model card ## `Jindo` is sLLM for construct datasets for LLM `KOLANI`. > Warning The training is still in progress. This model is an LLM in various language domains, including Korean translation and correction. Its main purpose is to create a dataset for training the Korean LLM "KOLANI" (which is still undergoing training). Furthermore, since this model has been developed solely by one individual without any external support, the release and improvement process might be relatively slow. Jindo is implemented as sLLM for lightweight purposes, thus focusing primarily on the 7B version for optimization, alongside the 13B version. Using: [QLoRA](https://github.com/artidoro/qlora) ## Model Details The weights you are currently viewing are preliminary checkpoints, and the official weights have not been released yet. * Developed by: [Minwoo Park](https://github.com/dsdanielpark) * Backbone Model: [LLaMA2](https://huggingface.co/meta-llama/Llama-2-7b) [[Paper](https://huggingface.co/papers/2307.09288)] * Model Jindo Variations: jindo-instruct * jindo-instruct Variations: 2b / 7b / 13b * [danielpark/ko-llama-2-jindo-2b-instruct]() (from LLaMA1) * [danielpark/ko-llama-2-jindo-7b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) (from LLaMA2) * [danielpark/ko-llama-2-jindo-13b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-13b-instruct) (from LLaMA2) * This model targets specific domains, so the 70b model will not be released. * Quantinized Weight: 7b-gptq(4bit-128g) * [ko-llama-2-jindo-7b-instruct-4bit-128g-gptq](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq) * Library: [HuggingFace Transformers](https://github.com/huggingface/transformers) * License: This model is licensed under the Meta's [LLaMA2 license](https://github.com/facebookresearch/llama/blob/main/LICENSE). We plan to check the dataset's license along with the official release, but our primary goal is to aim for a commercial-use release by default. * Where to send comments: Instructions on how to provide feedback or comments on a model can be found by opening an issue in the [Hugging Face community's model repository](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) * Contact: For questions and comments about the model, please email to me [[email protected]](mailto:[email protected]) ## Web Demo I implement the web demo using several popular tools that allow us to rapidly create web UIs. \| model \| web ui \| quantinized \| \| --- \| --- \| --- \| \| danielpark/ko-llama-2-jindo-7b-instruct. \| using [gradio](https://github.com/dsdanielpark/gradio) on [colab](https://colab.research.google.com/drive/1zwR7rz6Ym53tofCGwZZU8y5K_t1r1qqo#scrollTo=p2xw_g80xMsD) \| - \| \| danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq \| using [text-generation-webui](https://github.com/oobabooga/text-generation-webui) on [colab](https://colab.research.google.com/drive/19ihYHsyg_5QFZ_A28uZNR_Z68E_09L4G) \| gptq \| \| danielpark/ko-llama-2-jindo-7b-instruct-ggml \| [koboldcpp-v1.38](https://github.com/LostRuins/koboldcpp/releases/tag/v1.38) \| ggml \| ## Dataset Details ### Used Datasets - korean-jindo-dataset - The dataset has not been released yet > No other data was used except for the dataset mentioned above ### Prompt Template ``` ### System: {System} ### User: {User} ### Assistant: {Assistant} ``` ## Hardware and Software * Hardware * Under 10b model: Trained using the free T4 GPU resource. * Over 10b model: Utilized a Single A100 on Google Colab. * Training Factors: [HuggingFace trainer](https://huggingface.co/docs/transformers/main_classes/trainer) ## Evaluation Results Please refer to the following procedure for the evaluation of the backbone model. Other benchmarking and qualitative evaluations for Korean datasets are still pending. ### Overview - We conducted a performance evaluation based on the tasks being evaluated on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). We evaluated our model on four benchmark datasets, which include `ARC-Challenge`, `HellaSwag`, `MMLU`, and `TruthfulQA`. We used the [lm-evaluation-harness repository](https://github.com/EleutherAI/lm-evaluation-harness), specifically commit [b281b0921b636bc36ad05c0b0b0763bd6dd43463](https://github.com/EleutherAI/lm-evaluation-harness/tree/b281b0921b636bc36ad05c0b0b0763bd6dd43463). ## Usage Please refer to the following information and install the appropriate versions compatible with your enviroments. ``` $ pip install -q accelerate==0.21.0 peft==0.4.0 bitsandbytes==0.40.2 transformers==4.31.0 trl==0.4.7 ``` ```python from transformers import AutoTokenizer import transformers import torch model = "danielpark/ko-llama-2-jindo-13b-instruct" # model = "meta-llama/Llama-2-13b-hf" tokenizer = AutoTokenizer.from_pretrained(model) pipeline = transformers.pipeline( "text-generation", model=model, torch_dtype=torch.float16, device_map="auto", ) sequences = pipeline( 'I liked "Breaking Bad" and "Band of Brothers". Do you have any recommendations of other shows I might like?\n', do_sample=True, top_k=10, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id, max_length=200, ) for seq in sequences: print(f"Result: {seq['generated_text']}") ``` To use the model with the `transformers` library on a machine with GPUs, first make sure you have the `transformers` and `accelerate` libraries installed. ```python %pip install "accelerate>=0.16.0,<1" "transformers[torch]>=4.28.1,<5" "torch>=1.13.1,<2" ``` The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline` found in the model repo [here](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required. Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality. It is also fine to remove it if there is sufficient memory. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-13b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto") ``` You can then use the pipeline to answer instructions: ```python res = generate_text("Explain to me the difference between nuclear fission and fusion.") print(res[0]["generated_text"]) ``` Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer: ```python import torch from instruct_pipeline import InstructionTextGenerationPipeline from transformers import AutoModelForCausalLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", padding_side="left") model = AutoModelForCausalLM.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", device_map="auto", torch_dtype=torch.bfloat16) generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer) ``` ### LangChain Usage To use the pipeline with LangChain, you must set `return_full_text=True`, as LangChain expects the full text to be returned and the default for the pipeline is to only return the new text. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-7b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto", return_full_text=True) ``` You can create a prompt that either has only an instruction or has an instruction with context: ```python from langchain import PromptTemplate, LLMChain from langchain.llms import HuggingFacePipeline # template for an instrution with no input prompt = PromptTemplate( input_variables=["instruction"], template="{instruction}") # template for an instruction with input prompt_with_context = PromptTemplate( input_variables=["instruction", "context"], template="{instruction}\n\nInput:\n{context}") hf_pipeline = HuggingFacePipeline(pipeline=generate_text) llm_chain = LLMChain(llm=hf_pipeline, prompt=prompt) llm_context_chain = LLMChain(llm=hf_pipeline, prompt=prompt_with_context) ``` Example predicting using a simple instruction: ```python print(llm_chain.predict(instruction="Explain to me the difference between nuclear fission and fusion.").lstrip()) ``` Example predicting using an instruction with context: ```python context = """George Washington (February 22, 1732[b] - December 14, 1799) was an American military officer, statesman, and Founding Father who served as the first president of the United States from 1789 to 1797.""" print(llm_context_chain.predict(instruction="When was George Washington president?", context=context).lstrip()) ``` ### Scripts - Prepare evaluation environments: ``` # clone the repository git clone https://github.com/EleutherAI/lm-evaluation-harness.git # check out the specific commit git checkout b281b0921b636bc36ad05c0b0b0763bd6dd43463 # change to the repository directory cd lm-evaluation-harness ``` ## Ethical Issues The Jindo model has not been filtered for harmful, biased, or explicit content. As a result, outputs that do not adhere to ethical norms may be generated during use. Please exercise caution when using the model in research or practical applications. ### Ethical Considerations - There were no ethical issues involved, as we did not include the benchmark test set or the training set in the model's training process. As always, we encourage responsible and ethical use of this model. Please note that while Guanaco strives to provide accurate and helpful responses, it is still crucial to cross-verify the information from reliable sources for knowledge-based queries. ## Contact Me To contact me, you can [mail to me. [email protected]](mailto:[email protected]). ## Model Architecture ```python LlamaForCausalLM( (model): LlamaModel( (embed_tokens): Embedding(32000, 4096, padding_idx=0) (layers): ModuleList( (0-31): 32 x LlamaDecoderLayer( (self_attn): LlamaAttention( (q_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (k_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (v_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (o_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (rotary_emb): LlamaRotaryEmbedding() ) (mlp): LlamaMLP( (gate_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (up_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (down_proj): Linear4bit(in_features=11008, out_features=4096, bias=False) (act_fn): SiLUActivation() ) (input_layernorm): LlamaRMSNorm() (post_attention_layernorm): LlamaRMSNorm() ) ) (norm): LlamaRMSNorm() ) (lm_head): Linear(in_features=4096, out_features=32000, bias=False) ) ``` ## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0 ### License: The licenses of the pretrained models, llama1, and llama2, along with the datasets used, are applicable. For other datasets related to this work, guidance will be provided in the official release. The responsibility for verifying all licenses lies with the user, and the developer assumes no liability, explicit or implied, including legal responsibilities. ### Remark: - The "instruct" in the model name can be omitted, but it is used to differentiate between the backbones of llama2 for chat and general purposes. Additionally, this model is created for a specific purpose, so we plan to fine-tune it with a dataset focused on instructions. ### Naive Cost Estimation Assuming linearity without considering various variables. - 1000 simple prompts - 20 minutes of processing time - Approximately $2 in cost (based on Google Colab's 100 computing units and a single A100 GPU estimated at $10). - Required CPU RAM: 5GB (depending on the training data and dummy size) - Required VRAM: 12-13GB Time (\(t\)) and cost (\(c\)) are calculated based on the given information as follows: \[ t(n) = \frac{{20 \text{ minutes}}}{{1000 \text{ prompts}}} \times n \] \[ c(n) = \frac{{2 \text{ dollars}}}{{1000 \text{ prompts}}} \times n \] ```python def calculate_time_cost(num_prompts): total_prompts = 1000 total_time_minutes = 20 total_cost_dollars = 2 time_required = (total_time_minutes / total_prompts) * num_prompts cost_required = (total_cost_dollars / total_prompts) * num_prompts return time_required, cost_required # Example num_prompts = n time, cost = calculate_time_cost(num_prompts) print(f"Time for {num_prompts}: {time} minutes") print(f"Cost for {num_prompts}: {cost} dollar") ``` ### Chinchilla scaling laws The Chinchilla scaling laws focus on optimally scaling training compute but often we also care about inference cost. This tool follows [Harm de Vries’ blog post](https://www.harmdevries.com/post/model-size-vs-compute-overhead/) and visualizes the tradeoff between training comput and inference cost (i.e. model size).	null	Non_BioNLP	# Since this model is still under development, I recommend not using it until it reaches the development stage 5. Development Status :: 2 - Pre-Alpha <br> Developed by MinWoo Park, 2023, Seoul, South Korea. [Contact: [email protected]](mailto:[email protected]). [![Hits](https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fhuggingface.co%2Fdanielpark%2Fko-llama-2-jindo-13b-instruct&count_bg=%23000000&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=views&edge_flat=false)](https://hits.seeyoufarm.com) # danielpark/llama2-jindo-13b-instruct model card ## `Jindo` is sLLM for construct datasets for LLM `KOLANI`. > Warning The training is still in progress. This model is an LLM in various language domains, including Korean translation and correction. Its main purpose is to create a dataset for training the Korean LLM "KOLANI" (which is still undergoing training). Furthermore, since this model has been developed solely by one individual without any external support, the release and improvement process might be relatively slow. Jindo is implemented as sLLM for lightweight purposes, thus focusing primarily on the 7B version for optimization, alongside the 13B version. Using: [QLoRA](https://github.com/artidoro/qlora) ## Model Details The weights you are currently viewing are preliminary checkpoints, and the official weights have not been released yet. * Developed by: [Minwoo Park](https://github.com/dsdanielpark) * Backbone Model: [LLaMA2](https://huggingface.co/meta-llama/Llama-2-7b) [[Paper](https://huggingface.co/papers/2307.09288)] * Model Jindo Variations: jindo-instruct * jindo-instruct Variations: 2b / 7b / 13b * [danielpark/ko-llama-2-jindo-2b-instruct]() (from LLaMA1) * [danielpark/ko-llama-2-jindo-7b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) (from LLaMA2) * [danielpark/ko-llama-2-jindo-13b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-13b-instruct) (from LLaMA2) * This model targets specific domains, so the 70b model will not be released. * Quantinized Weight: 7b-gptq(4bit-128g) * [ko-llama-2-jindo-7b-instruct-4bit-128g-gptq](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq) * Library: [HuggingFace Transformers](https://github.com/huggingface/transformers) * License: This model is licensed under the Meta's [LLaMA2 license](https://github.com/facebookresearch/llama/blob/main/LICENSE). We plan to check the dataset's license along with the official release, but our primary goal is to aim for a commercial-use release by default. * Where to send comments: Instructions on how to provide feedback or comments on a model can be found by opening an issue in the [Hugging Face community's model repository](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) * Contact: For questions and comments about the model, please email to me [[email protected]](mailto:[email protected]) ## Web Demo I implement the web demo using several popular tools that allow us to rapidly create web UIs. \| model \| web ui \| quantinized \| \| --- \| --- \| --- \| \| danielpark/ko-llama-2-jindo-7b-instruct. \| using [gradio](https://github.com/dsdanielpark/gradio) on [colab](https://colab.research.google.com/drive/1zwR7rz6Ym53tofCGwZZU8y5K_t1r1qqo#scrollTo=p2xw_g80xMsD) \| - \| \| danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq \| using [text-generation-webui](https://github.com/oobabooga/text-generation-webui) on [colab](https://colab.research.google.com/drive/19ihYHsyg_5QFZ_A28uZNR_Z68E_09L4G) \| gptq \| \| danielpark/ko-llama-2-jindo-7b-instruct-ggml \| [koboldcpp-v1.38](https://github.com/LostRuins/koboldcpp/releases/tag/v1.38) \| ggml \| ## Dataset Details ### Used Datasets - korean-jindo-dataset - The dataset has not been released yet > No other data was used except for the dataset mentioned above ### Prompt Template ``` ### System: {System} ### User: {User} ### Assistant: {Assistant} ``` ## Hardware and Software * Hardware * Under 10b model: Trained using the free T4 GPU resource. * Over 10b model: Utilized a Single A100 on Google Colab. * Training Factors: [HuggingFace trainer](https://huggingface.co/docs/transformers/main_classes/trainer) ## Evaluation Results Please refer to the following procedure for the evaluation of the backbone model. Other benchmarking and qualitative evaluations for Korean datasets are still pending. ### Overview - We conducted a performance evaluation based on the tasks being evaluated on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). We evaluated our model on four benchmark datasets, which include `ARC-Challenge`, `HellaSwag`, `MMLU`, and `TruthfulQA`. We used the [lm-evaluation-harness repository](https://github.com/EleutherAI/lm-evaluation-harness), specifically commit [b281b0921b636bc36ad05c0b0b0763bd6dd43463](https://github.com/EleutherAI/lm-evaluation-harness/tree/b281b0921b636bc36ad05c0b0b0763bd6dd43463). ## Usage Please refer to the following information and install the appropriate versions compatible with your enviroments. ``` $ pip install -q accelerate==0.21.0 peft==0.4.0 bitsandbytes==0.40.2 transformers==4.31.0 trl==0.4.7 ``` ```python from transformers import AutoTokenizer import transformers import torch model = "danielpark/ko-llama-2-jindo-13b-instruct" # model = "meta-llama/Llama-2-13b-hf" tokenizer = AutoTokenizer.from_pretrained(model) pipeline = transformers.pipeline( "text-generation", model=model, torch_dtype=torch.float16, device_map="auto", ) sequences = pipeline( 'I liked "Breaking Bad" and "Band of Brothers". Do you have any recommendations of other shows I might like?\n', do_sample=True, top_k=10, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id, max_length=200, ) for seq in sequences: print(f"Result: {seq['generated_text']}") ``` To use the model with the `transformers` library on a machine with GPUs, first make sure you have the `transformers` and `accelerate` libraries installed. ```python %pip install "accelerate>=0.16.0,<1" "transformers[torch]>=4.28.1,<5" "torch>=1.13.1,<2" ``` The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline` found in the model repo [here](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required. Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality. It is also fine to remove it if there is sufficient memory. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-13b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto") ``` You can then use the pipeline to answer instructions: ```python res = generate_text("Explain to me the difference between nuclear fission and fusion.") print(res[0]["generated_text"]) ``` Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer: ```python import torch from instruct_pipeline import InstructionTextGenerationPipeline from transformers import AutoModelForCausalLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", padding_side="left") model = AutoModelForCausalLM.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", device_map="auto", torch_dtype=torch.bfloat16) generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer) ``` ### LangChain Usage To use the pipeline with LangChain, you must set `return_full_text=True`, as LangChain expects the full text to be returned and the default for the pipeline is to only return the new text. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-7b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto", return_full_text=True) ``` You can create a prompt that either has only an instruction or has an instruction with context: ```python from langchain import PromptTemplate, LLMChain from langchain.llms import HuggingFacePipeline # template for an instrution with no input prompt = PromptTemplate( input_variables=["instruction"], template="{instruction}") # template for an instruction with input prompt_with_context = PromptTemplate( input_variables=["instruction", "context"], template="{instruction}\n\nInput:\n{context}") hf_pipeline = HuggingFacePipeline(pipeline=generate_text) llm_chain = LLMChain(llm=hf_pipeline, prompt=prompt) llm_context_chain = LLMChain(llm=hf_pipeline, prompt=prompt_with_context) ``` Example predicting using a simple instruction: ```python print(llm_chain.predict(instruction="Explain to me the difference between nuclear fission and fusion.").lstrip()) ``` Example predicting using an instruction with context: ```python context = """George Washington (February 22, 1732[b] - December 14, 1799) was an American military officer, statesman, and Founding Father who served as the first president of the United States from 1789 to 1797.""" print(llm_context_chain.predict(instruction="When was George Washington president?", context=context).lstrip()) ``` ### Scripts - Prepare evaluation environments: ``` # clone the repository git clone https://github.com/EleutherAI/lm-evaluation-harness.git # check out the specific commit git checkout b281b0921b636bc36ad05c0b0b0763bd6dd43463 # change to the repository directory cd lm-evaluation-harness ``` ## Ethical Issues The Jindo model has not been filtered for harmful, biased, or explicit content. As a result, outputs that do not adhere to ethical norms may be generated during use. Please exercise caution when using the model in research or practical applications. ### Ethical Considerations - There were no ethical issues involved, as we did not include the benchmark test set or the training set in the model's training process. As always, we encourage responsible and ethical use of this model. Please note that while Guanaco strives to provide accurate and helpful responses, it is still crucial to cross-verify the information from reliable sources for knowledge-based queries. ## Contact Me To contact me, you can [mail to me. [email protected]](mailto:[email protected]). ## Model Architecture ```python LlamaForCausalLM( (model): LlamaModel( (embed_tokens): Embedding(32000, 4096, padding_idx=0) (layers): ModuleList( (0-31): 32 x LlamaDecoderLayer( (self_attn): LlamaAttention( (q_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (k_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (v_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (o_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (rotary_emb): LlamaRotaryEmbedding() ) (mlp): LlamaMLP( (gate_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (up_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (down_proj): Linear4bit(in_features=11008, out_features=4096, bias=False) (act_fn): SiLUActivation() ) (input_layernorm): LlamaRMSNorm() (post_attention_layernorm): LlamaRMSNorm() ) ) (norm): LlamaRMSNorm() ) (lm_head): Linear(in_features=4096, out_features=32000, bias=False) ) ``` ## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0 ### License: The licenses of the pretrained models, llama1, and llama2, along with the datasets used, are applicable. For other datasets related to this work, guidance will be provided in the official release. The responsibility for verifying all licenses lies with the user, and the developer assumes no liability, explicit or implied, including legal responsibilities. ### Remark: - The "instruct" in the model name can be omitted, but it is used to differentiate between the backbones of llama2 for chat and general purposes. Additionally, this model is created for a specific purpose, so we plan to fine-tune it with a dataset focused on instructions. ### Naive Cost Estimation Assuming linearity without considering various variables. - 1000 simple prompts - 20 minutes of processing time - Approximately $2 in cost (based on Google Colab's 100 computing units and a single A100 GPU estimated at $10). - Required CPU RAM: 5GB (depending on the training data and dummy size) - Required VRAM: 12-13GB Time (\(t\)) and cost (\(c\)) are calculated based on the given information as follows: \[ t(n) = \frac{{20 \text{ minutes}}}{{1000 \text{ prompts}}} \times n \] \[ c(n) = \frac{{2 \text{ dollars}}}{{1000 \text{ prompts}}} \times n \] ```python def calculate_time_cost(num_prompts): total_prompts = 1000 total_time_minutes = 20 total_cost_dollars = 2 time_required = (total_time_minutes / total_prompts) * num_prompts cost_required = (total_cost_dollars / total_prompts) * num_prompts return time_required, cost_required # Example num_prompts = n time, cost = calculate_time_cost(num_prompts) print(f"Time for {num_prompts}: {time} minutes") print(f"Cost for {num_prompts}: {cost} dollar") ``` ### Chinchilla scaling laws The Chinchilla scaling laws focus on optimally scaling training compute but often we also care about inference cost. This tool follows [Harm de Vries’ blog post](https://www.harmdevries.com/post/model-size-vs-compute-overhead/) and visualizes the tradeoff between training comput and inference cost (i.e. model size).	{"datasets": ["korean-jindo-dataset.json"], "language": ["en", "ko"], "library_name": "peft", "pipeline_tag": "text-generation", "tags": ["dsdanielpark", "llama2", "instruct", "instruction", "jindo", "korean", "translation", "13b"]}	task	[ "TRANSLATION" ]	44,599
cahya/bert2bert-indonesian-summarization	cahya	summarization	[ "transformers", "pytorch", "encoder-decoder", "text2text-generation", "pipeline:summarization", "summarization", "bert2bert", "id", "dataset:id_liputan6", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:05Z	2021-01-29T11:39:42+00:00	480	4	--- datasets: - id_liputan6 language: id license: apache-2.0 tags: - pipeline:summarization - summarization - bert2bert --- # Indonesian BERT2BERT Summarization Model Finetuned BERT-base summarization model for Indonesian. ## Finetuning Corpus `bert2bert-indonesian-summarization` model is based on `cahya/bert-base-indonesian-1.5G` by [cahya](https://huggingface.co/cahya), finetuned using [id_liputan6](https://huggingface.co/datasets/id_liputan6) dataset. ## Load Finetuned Model ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") ``` ## Code Sample ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") # ARTICLE_TO_SUMMARIZE = "" # generate summary input_ids = tokenizer.encode(ARTICLE_TO_SUMMARIZE, return_tensors='pt') summary_ids = model.generate(input_ids, min_length=20, max_length=80, num_beams=10, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True, no_repeat_ngram_size=2, use_cache=True, do_sample = True, temperature = 0.8, top_k = 50, top_p = 0.95) summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True) print(summary_text) ``` Output: ``` ```	null	Non_BioNLP	# Indonesian BERT2BERT Summarization Model Finetuned BERT-base summarization model for Indonesian. ## Finetuning Corpus `bert2bert-indonesian-summarization` model is based on `cahya/bert-base-indonesian-1.5G` by [cahya](https://huggingface.co/cahya), finetuned using [id_liputan6](https://huggingface.co/datasets/id_liputan6) dataset. ## Load Finetuned Model ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") ``` ## Code Sample ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") # ARTICLE_TO_SUMMARIZE = "" # generate summary input_ids = tokenizer.encode(ARTICLE_TO_SUMMARIZE, return_tensors='pt') summary_ids = model.generate(input_ids, min_length=20, max_length=80, num_beams=10, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True, no_repeat_ngram_size=2, use_cache=True, do_sample = True, temperature = 0.8, top_k = 50, top_p = 0.95) summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True) print(summary_text) ``` Output: ``` ```	{"datasets": ["id_liputan6"], "language": "id", "license": "apache-2.0", "tags": ["pipeline:summarization", "summarization", "bert2bert"]}	task	[ "SUMMARIZATION" ]	44,600
m3hrdadfi/xlmr-large-qa-sv	m3hrdadfi	question-answering	[ "transformers", "pytorch", "tf", "xlm-roberta", "question-answering", "roberta", "squad", "sv", "multilingual", "model-index", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:05Z	2021-10-12T13:50:27+00:00	111	2	--- language: - sv - multilingual metrics: - squad_v2 tags: - question-answering - xlm-roberta - roberta - squad widget: - text: Vilket datum är den svenska nationaldagen? context: Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. - text: Vad innebär helgdag i Sverige? context: Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. - text: Vilket år tillkom Sveriges nationaldag? context: Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. model-index: - name: XLM-RoBERTa large for QA (SwedishQA - 🇸🇪) results: - task: type: question-answering name: Question Answering dataset: name: SwedishQA type: swedish_qa args: sv metrics: - type: squad_v2 value: 87.97 name: Eval F1 args: max_order - type: squad_v2 value: 78.79 name: Eval Exact args: max_order --- # XLM-RoBERTa large for QA (SwedishQA - 🇸🇪) This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the [SwedishQA](https://github.com/Vottivott/building-a-swedish-qa-model) dataset. ## Hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 2.0 - mixed_precision_training: Native AMP ## Performance Evaluation results on the eval set with the official [eval script](https://worksheets.codalab.org/rest/bundles/0x6b567e1cf2e041ec80d7098f031c5c9e/contents/blob/). ### Evalset ```text "exact": 78.79554655870446, "f1": 87.97339064752278, "total": 5928 ``` ## Usage ```python from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline model_name_or_path = "m3hrdadfi/xlmr-large-qa-sv" nlp = pipeline('question-answering', model=model_name_or_path, tokenizer=model_name_or_path) context = """ Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. """ questions = [ "Vilket datum är den svenska nationaldagen?", "Vad innebär helgdag i Sverige?", "Vilket år tillkom Sveriges nationaldag?" ] kwargs = {} for question in questions: r = nlp(question=question, context=context, kwargs) answer = " ".join([token.strip() for token in r["answer"].strip().split() if token.strip()]) print(f"{question} {answer}") ``` Output** ```text Vilket datum är den svenska nationaldagen? 6 juni Vad innebär helgdag i Sverige? svenska flaggans dag Vilket år tillkom Sveriges nationaldag? 1983 ``` ## Authors - [Mehrdad Farahani](https://github.com/m3hrdadfi) ### Framework versions - Transformers 4.12.0.dev0 - Pytorch 1.9.1+cu111 - Datasets 1.12.1 - Tokenizers 0.10.3	null	Non_BioNLP	# XLM-RoBERTa large for QA (SwedishQA - 🇸🇪) This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the [SwedishQA](https://github.com/Vottivott/building-a-swedish-qa-model) dataset. ## Hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 2.0 - mixed_precision_training: Native AMP ## Performance Evaluation results on the eval set with the official [eval script](https://worksheets.codalab.org/rest/bundles/0x6b567e1cf2e041ec80d7098f031c5c9e/contents/blob/). ### Evalset ```text "exact": 78.79554655870446, "f1": 87.97339064752278, "total": 5928 ``` ## Usage ```python from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline model_name_or_path = "m3hrdadfi/xlmr-large-qa-sv" nlp = pipeline('question-answering', model=model_name_or_path, tokenizer=model_name_or_path) context = """ Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. """ questions = [ "Vilket datum är den svenska nationaldagen?", "Vad innebär helgdag i Sverige?", "Vilket år tillkom Sveriges nationaldag?" ] kwargs = {} for question in questions: r = nlp(question=question, context=context, kwargs) answer = " ".join([token.strip() for token in r["answer"].strip().split() if token.strip()]) print(f"{question} {answer}") ``` Output** ```text Vilket datum är den svenska nationaldagen? 6 juni Vad innebär helgdag i Sverige? svenska flaggans dag Vilket år tillkom Sveriges nationaldag? 1983 ``` ## Authors - [Mehrdad Farahani](https://github.com/m3hrdadfi) ### Framework versions - Transformers 4.12.0.dev0 - Pytorch 1.9.1+cu111 - Datasets 1.12.1 - Tokenizers 0.10.3	{"language": ["sv", "multilingual"], "metrics": ["squad_v2"], "tags": ["question-answering", "xlm-roberta", "roberta", "squad"], "widget": [{"text": "Vilket datum är den svenska nationaldagen?", "context": "Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som \"svenska flaggans dag\" och det var först 1983 som dagen även fick status som nationaldag."}, {"text": "Vad innebär helgdag i Sverige?", "context": "Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som \"svenska flaggans dag\" och det var först 1983 som dagen även fick status som nationaldag."}, {"text": "Vilket år tillkom Sveriges nationaldag?", "context": "Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som \"svenska flaggans dag\" och det var först 1983 som dagen även fick status som nationaldag."}], "model-index": [{"name": "XLM-RoBERTa large for QA (SwedishQA - 🇸🇪)", "results": [{"task": {"type": "question-answering", "name": "Question Answering"}, "dataset": {"name": "SwedishQA", "type": "swedish_qa", "args": "sv"}, "metrics": [{"type": "squad_v2", "value": 87.97, "name": "Eval F1", "args": "max_order"}, {"type": "squad_v2", "value": 78.79, "name": "Eval Exact", "args": "max_order"}]}]}]}	task	[ "QUESTION_ANSWERING" ]	44,601
Helsinki-NLP/opus-mt-swc-sv	Helsinki-NLP	translation	[ "transformers", "pytorch", "tf", "marian", "text2text-generation", "translation", "swc", "sv", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:04Z	2023-08-16T12:06:38+00:00	372	0	--- license: apache-2.0 tags: - translation --- ### opus-mt-swc-sv * source languages: swc * target languages: sv * OPUS readme: [swc-sv](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/swc-sv/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.zip) * test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.test.txt) * test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.eval.txt) ## Benchmarks \| testset \| BLEU \| chr-F \| \|-----------------------\|-------\|-------\| \| JW300.swc.sv \| 30.7 \| 0.495 \|	null	Non_BioNLP	### opus-mt-swc-sv * source languages: swc * target languages: sv * OPUS readme: [swc-sv](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/swc-sv/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.zip) * test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.test.txt) * test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.eval.txt) ## Benchmarks \| testset \| BLEU \| chr-F \| \|-----------------------\|-------\|-------\| \| JW300.swc.sv \| 30.7 \| 0.495 \|	{"license": "apache-2.0", "tags": ["translation"]}	task	[ "TRANSLATION" ]	44,602
mlninja-dev/distilbert-base-uncased-finetuned-emotion	mlninja-dev	text-classification	[ "transformers", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "dataset:emotion", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-11-10T07:43:27Z	2023-11-10T08:13:53+00:00	8	0	--- base_model: distilbert-base-uncased datasets: - emotion license: apache-2.0 metrics: - accuracy - f1 - precision - recall tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion config: split split: validation args: split metrics: - type: accuracy value: 0.9265 name: Accuracy - type: f1 value: 0.9265091764064305 name: F1 - type: precision value: 0.9265685739545657 name: Precision - type: recall value: 0.9265 name: Recall --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2114 - Accuracy: 0.9265 - F1: 0.9265 - Precision: 0.9266 - Recall: 0.9265 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| Precision \| Recall \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\|:---------:\|:------:\| \| 0.8163 \| 1.0 \| 250 \| 0.3191 \| 0.9055 \| 0.9045 \| 0.9083 \| 0.9055 \| \| 0.2509 \| 2.0 \| 500 \| 0.2114 \| 0.9265 \| 0.9265 \| 0.9266 \| 0.9265 \| ### Framework versions - Transformers 4.35.0 - Pytorch 2.1.0 - Datasets 2.14.6 - Tokenizers 0.14.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2114 - Accuracy: 0.9265 - F1: 0.9265 - Precision: 0.9266 - Recall: 0.9265 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| Precision \| Recall \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\|:---------:\|:------:\| \| 0.8163 \| 1.0 \| 250 \| 0.3191 \| 0.9055 \| 0.9045 \| 0.9083 \| 0.9055 \| \| 0.2509 \| 2.0 \| 500 \| 0.2114 \| 0.9265 \| 0.9265 \| 0.9266 \| 0.9265 \| ### Framework versions - Transformers 4.35.0 - Pytorch 2.1.0 - Datasets 2.14.6 - Tokenizers 0.14.1	{"base_model": "distilbert-base-uncased", "datasets": ["emotion"], "license": "apache-2.0", "metrics": ["accuracy", "f1", "precision", "recall"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-emotion", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "config": "split", "split": "validation", "args": "split"}, "metrics": [{"type": "accuracy", "value": 0.9265, "name": "Accuracy"}, {"type": "f1", "value": 0.9265091764064305, "name": "F1"}, {"type": "precision", "value": 0.9265685739545657, "name": "Precision"}, {"type": "recall", "value": 0.9265, "name": "Recall"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,603
prithivMLmods/Llama-Chat-Summary-3.2-3B-GGUF	prithivMLmods	text-generation	[ "transformers", "gguf", "llama", "safetensors", "ollama", "llama-cpp", "text-generation-inference", "chat-summary", "text-generation", "en", "dataset:prithivMLmods/Context-Based-Chat-Summary-Plus", "base_model:prithivMLmods/Llama-Chat-Summary-3.2-3B", "base_model:quantized:prithivMLmods/Llama-Chat-Summary-3.2-3B", "license:creativeml-openrail-m", "endpoints_compatible", "region:us", "conversational" ]	2024-12-17T04:34:28Z	2024-12-17T09:46:03+00:00	941	3	--- base_model: - prithivMLmods/Llama-Chat-Summary-3.2-3B datasets: - prithivMLmods/Context-Based-Chat-Summary-Plus language: - en library_name: transformers license: creativeml-openrail-m pipeline_tag: text-generation tags: - safetensors - ollama - llama-cpp - text-generation-inference - chat-summary --- ### Llama-Chat-Summary-3.2-3B: Context-Aware Summarization Model Llama-Chat-Summary-3.2-3B is a fine-tuned model designed for generating context-aware summaries of long conversational or text-based inputs. Built on the meta-llama/Llama-3.2-3B-Instruct foundation, this model is optimized to process structured and unstructured conversational data for summarization tasks. \| File Name \| Size \| Description \| Upload Status \| \|--------------------------------------------\|------------------\|--------------------------------------------------\|-------------------\| \| `.gitattributes` \| 1.81 kB \| Git LFS tracking configuration. \| Uploaded \| \| `Llama-Chat-Summary-3.2-3B.F16.gguf` \| 6.43 GB \| Full precision (F16) GGUF model file. \| Uploaded (LFS) \| \| `Llama-Chat-Summary-3.2-3B.Q4_K_M.gguf` \| 2.02 GB \| Quantized Q4_K_M GGUF model file. \| Uploaded (LFS) \| \| `Llama-Chat-Summary-3.2-3B.Q5_K_M.gguf` \| 2.32 GB \| Quantized Q5_K_M GGUF model file. \| Uploaded (LFS) \| \| `Llama-Chat-Summary-3.2-3B.Q8_0.gguf` \| 3.42 GB \| Quantized Q8_0 GGUF model file. \| Uploaded (LFS) \| \| `Modelfile` \| 2.03 kB \| Model configuration or build script file. \| Uploaded \| \| `README.md` \| 42 Bytes \| Minimal commit message placeholder. \| Uploaded \| \| `config.json` \| 31 Bytes \| Model metadata and configuration. \| Uploaded \| ### Key Features 1. Conversation Summarization: - Generates concise and meaningful summaries of long chats, discussions, or threads. 2. Context Preservation: - Maintains critical points, ensuring important details aren't omitted. 3. Text Summarization: - Works beyond chats; supports summarizing articles, documents, or reports. 4. Fine-Tuned Efficiency: - Trained with Context-Based-Chat-Summary-Plus dataset for accurate summarization of chat and conversational data. --- ### Training Details - Base Model: [meta-llama/Llama-3.2-3B-Instruct](#) - Fine-Tuning Dataset: [prithivMLmods/Context-Based-Chat-Summary-Plus](#) - Contains 98.4k structured and unstructured conversations, summaries, and contextual inputs for robust training. --- ### Applications 1. Customer Support Logs: - Summarize chat logs or support tickets for insights and reporting. 2. Meeting Notes: - Generate concise summaries of meeting transcripts. 3. Document Summarization: - Create short summaries for lengthy reports or articles. 4. Content Generation Pipelines: - Automate summarization for newsletters, blogs, or email digests. 5. Context Extraction for AI Systems: - Preprocess chat or conversation logs for downstream AI applications. #### Load the Model ```python from transformers import AutoModelForCausalLM, AutoTokenizer model_name = "prithivMLmods/Llama-Chat-Summary-3.2-3B" tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained(model_name) ``` #### Generate a Summary ```python prompt = """ Summarize the following conversation: User1: Hey, I need help with my order. It hasn't arrived yet. User2: I'm sorry to hear that. Can you provide your order number? User1: Sure, it's 12345. User2: Let me check... It seems there was a delay. It should arrive tomorrow. User1: Okay, thank you! """ inputs = tokenizer(prompt, return_tensors="pt") outputs = model.generate(inputs, max_length=100, temperature=0.7) summary = tokenizer.decode(outputs[0], skip_special_tokens=True) print("Summary:", summary) ``` --- ### Expected Output "The user reported a delayed order (12345), and support confirmed it will arrive tomorrow." --- ### Deployment Notes - Serverless API: This model currently lacks sufficient usage for serverless endpoints. Use dedicated endpoints for deployment. - Performance Requirements:** - GPU with sufficient memory (recommended for large models). - Optimization techniques like quantization can improve efficiency for inference. --- # Run with Ollama [ Ollama Run ] ## Overview Ollama is a powerful tool that allows you to run machine learning models effortlessly. This guide will help you download, install, and run your own GGUF models in just a few minutes. ## Table of Contents - [Download and Install Ollama](#download-and-install-ollama) - [Steps to Run GGUF Models](#steps-to-run-gguf-models) - [1. Create the Model File](#1-create-the-model-file) - [2. Add the Template Command](#2-add-the-template-command) - [3. Create and Patch the Model](#3-create-and-patch-the-model) - [Running the Model](#running-the-model) - [Sample Usage](#sample-usage) ## Download and Install Ollama🦙 To get started, download Ollama from [https://ollama.com/download](https://ollama.com/download) and install it on your Windows or Mac system. ## Steps to Run GGUF Models ### 1. Create the Model File First, create a model file and name it appropriately. For example, you can name your model file `metallama`. ### 2. Add the Template Command In your model file, include a `FROM` line that specifies the base model file you want to use. For instance: ```bash FROM Llama-3.2-1B.F16.gguf ``` Ensure that the model file is in the same directory as your script. ### 3. Create and Patch the Model Open your terminal and run the following command to create and patch your model: ```bash ollama create metallama -f ./metallama ``` Once the process is successful, you will see a confirmation message. To verify that the model was created successfully, you can list all models with: ```bash ollama list ``` Make sure that `metallama` appears in the list of models. --- ## Running the Model To run your newly created model, use the following command in your terminal: ```bash ollama run metallama ``` ### Sample Usage / Test In the command prompt, you can execute: ```bash D:\>ollama run metallama ``` You can interact with the model like this: ```plaintext >>> write a mini passage about space x Space X, the private aerospace company founded by Elon Musk, is revolutionizing the field of space exploration. With its ambitious goals to make humanity a multi-planetary species and establish a sustainable human presence in the cosmos, Space X has become a leading player in the industry. The company's spacecraft, like the Falcon 9, have demonstrated remarkable capabilities, allowing for the transport of crews and cargo into space with unprecedented efficiency. As technology continues to advance, the possibility of establishing permanent colonies on Mars becomes increasingly feasible, thanks in part to the success of reusable rockets that can launch multiple times without sustaining significant damage. The journey towards becoming a multi-planetary species is underway, and Space X plays a pivotal role in pushing the boundaries of human exploration and settlement. ``` --- ## Conclusion With these simple steps, you can easily download, install, and run your own models using Ollama. Whether you're exploring the capabilities of Llama or building your own custom models, Ollama makes it accessible and efficient. - This README provides clear instructions and structured information to help users navigate the process of using Ollama effectively. Adjust any sections as needed based on your specific requirements or additional details you may want to include. ---	null	Non_BioNLP	### Llama-Chat-Summary-3.2-3B: Context-Aware Summarization Model Llama-Chat-Summary-3.2-3B is a fine-tuned model designed for generating context-aware summaries of long conversational or text-based inputs. Built on the meta-llama/Llama-3.2-3B-Instruct foundation, this model is optimized to process structured and unstructured conversational data for summarization tasks. \| File Name \| Size \| Description \| Upload Status \| \|--------------------------------------------\|------------------\|--------------------------------------------------\|-------------------\| \| `.gitattributes` \| 1.81 kB \| Git LFS tracking configuration. \| Uploaded \| \| `Llama-Chat-Summary-3.2-3B.F16.gguf` \| 6.43 GB \| Full precision (F16) GGUF model file. \| Uploaded (LFS) \| \| `Llama-Chat-Summary-3.2-3B.Q4_K_M.gguf` \| 2.02 GB \| Quantized Q4_K_M GGUF model file. \| Uploaded (LFS) \| \| `Llama-Chat-Summary-3.2-3B.Q5_K_M.gguf` \| 2.32 GB \| Quantized Q5_K_M GGUF model file. \| Uploaded (LFS) \| \| `Llama-Chat-Summary-3.2-3B.Q8_0.gguf` \| 3.42 GB \| Quantized Q8_0 GGUF model file. \| Uploaded (LFS) \| \| `Modelfile` \| 2.03 kB \| Model configuration or build script file. \| Uploaded \| \| `README.md` \| 42 Bytes \| Minimal commit message placeholder. \| Uploaded \| \| `config.json` \| 31 Bytes \| Model metadata and configuration. \| Uploaded \| ### Key Features 1. Conversation Summarization: - Generates concise and meaningful summaries of long chats, discussions, or threads. 2. Context Preservation: - Maintains critical points, ensuring important details aren't omitted. 3. Text Summarization: - Works beyond chats; supports summarizing articles, documents, or reports. 4. Fine-Tuned Efficiency: - Trained with Context-Based-Chat-Summary-Plus dataset for accurate summarization of chat and conversational data. --- ### Training Details - Base Model: [meta-llama/Llama-3.2-3B-Instruct](#) - Fine-Tuning Dataset: [prithivMLmods/Context-Based-Chat-Summary-Plus](#) - Contains 98.4k structured and unstructured conversations, summaries, and contextual inputs for robust training. --- ### Applications 1. Customer Support Logs: - Summarize chat logs or support tickets for insights and reporting. 2. Meeting Notes: - Generate concise summaries of meeting transcripts. 3. Document Summarization: - Create short summaries for lengthy reports or articles. 4. Content Generation Pipelines: - Automate summarization for newsletters, blogs, or email digests. 5. Context Extraction for AI Systems: - Preprocess chat or conversation logs for downstream AI applications. #### Load the Model ```python from transformers import AutoModelForCausalLM, AutoTokenizer model_name = "prithivMLmods/Llama-Chat-Summary-3.2-3B" tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained(model_name) ``` #### Generate a Summary ```python prompt = """ Summarize the following conversation: User1: Hey, I need help with my order. It hasn't arrived yet. User2: I'm sorry to hear that. Can you provide your order number? User1: Sure, it's 12345. User2: Let me check... It seems there was a delay. It should arrive tomorrow. User1: Okay, thank you! """ inputs = tokenizer(prompt, return_tensors="pt") outputs = model.generate(inputs, max_length=100, temperature=0.7) summary = tokenizer.decode(outputs[0], skip_special_tokens=True) print("Summary:", summary) ``` --- ### Expected Output "The user reported a delayed order (12345), and support confirmed it will arrive tomorrow." --- ### Deployment Notes - Serverless API: This model currently lacks sufficient usage for serverless endpoints. Use dedicated endpoints for deployment. - Performance Requirements:** - GPU with sufficient memory (recommended for large models). - Optimization techniques like quantization can improve efficiency for inference. --- # Run with Ollama [ Ollama Run ] ## Overview Ollama is a powerful tool that allows you to run machine learning models effortlessly. This guide will help you download, install, and run your own GGUF models in just a few minutes. ## Table of Contents - [Download and Install Ollama](#download-and-install-ollama) - [Steps to Run GGUF Models](#steps-to-run-gguf-models) - [1. Create the Model File](#1-create-the-model-file) - [2. Add the Template Command](#2-add-the-template-command) - [3. Create and Patch the Model](#3-create-and-patch-the-model) - [Running the Model](#running-the-model) - [Sample Usage](#sample-usage) ## Download and Install Ollama🦙 To get started, download Ollama from [https://ollama.com/download](https://ollama.com/download) and install it on your Windows or Mac system. ## Steps to Run GGUF Models ### 1. Create the Model File First, create a model file and name it appropriately. For example, you can name your model file `metallama`. ### 2. Add the Template Command In your model file, include a `FROM` line that specifies the base model file you want to use. For instance: ```bash FROM Llama-3.2-1B.F16.gguf ``` Ensure that the model file is in the same directory as your script. ### 3. Create and Patch the Model Open your terminal and run the following command to create and patch your model: ```bash ollama create metallama -f ./metallama ``` Once the process is successful, you will see a confirmation message. To verify that the model was created successfully, you can list all models with: ```bash ollama list ``` Make sure that `metallama` appears in the list of models. --- ## Running the Model To run your newly created model, use the following command in your terminal: ```bash ollama run metallama ``` ### Sample Usage / Test In the command prompt, you can execute: ```bash D:\>ollama run metallama ``` You can interact with the model like this: ```plaintext >>> write a mini passage about space x Space X, the private aerospace company founded by Elon Musk, is revolutionizing the field of space exploration. With its ambitious goals to make humanity a multi-planetary species and establish a sustainable human presence in the cosmos, Space X has become a leading player in the industry. The company's spacecraft, like the Falcon 9, have demonstrated remarkable capabilities, allowing for the transport of crews and cargo into space with unprecedented efficiency. As technology continues to advance, the possibility of establishing permanent colonies on Mars becomes increasingly feasible, thanks in part to the success of reusable rockets that can launch multiple times without sustaining significant damage. The journey towards becoming a multi-planetary species is underway, and Space X plays a pivotal role in pushing the boundaries of human exploration and settlement. ``` --- ## Conclusion With these simple steps, you can easily download, install, and run your own models using Ollama. Whether you're exploring the capabilities of Llama or building your own custom models, Ollama makes it accessible and efficient. - This README provides clear instructions and structured information to help users navigate the process of using Ollama effectively. Adjust any sections as needed based on your specific requirements or additional details you may want to include. ---	{"base_model": ["prithivMLmods/Llama-Chat-Summary-3.2-3B"], "datasets": ["prithivMLmods/Context-Based-Chat-Summary-Plus"], "language": ["en"], "library_name": "transformers", "license": "creativeml-openrail-m", "pipeline_tag": "text-generation", "tags": ["safetensors", "ollama", "llama-cpp", "text-generation-inference", "chat-summary"]}	task	[ "SUMMARIZATION" ]	44,604
cnmoro/ptt5-base-ptbr-summarization	cnmoro	summarization	[ "transformers", "pytorch", "t5", "text2text-generation", "summarization", "pt", "license:mit", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2023-11-09T14:33:11Z	2023-11-10T12:54:28+00:00	651	2	--- language: - pt license: mit tags: - summarization ---	null	Non_BioNLP		{"language": ["pt"], "license": "mit", "tags": ["summarization"]}	task	[ "SUMMARIZATION" ]	44,605
deinon-daemon/axolotl-13b-chat-qlora-dev	deinon-daemon	text-generation	[ "transformers", "pytorch", "llama", "text-generation", "en", "license:llama2", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2023-07-29T11:48:10Z	2023-07-29T14:38:16+00:00	10	0	--- language: en license: llama2 task_categories: - conversational - question-answering - zero-shot-answering - summarization - text-generation - text2text-generation pretty_name: Axolotl-13B-chat --- <img src="https://huggingface.co/deinon-daemon/axolotl-13b-chat-qlora-dev/resolve/main/axolotl_img/a14a0db4-9c8c-4283-8d24-21eb8f6210b1.png" alt="¡ Say hi to Axolotl !" style="float: left; margin-right: 10px;" /> Say hello to axolotl: a small-is-powerful instruct-tuned chat model! This is my second build ever in the fine tuning world. It was hacked in about 48hrs, and was executed entirely on one colab kernel for ~8-9hrs last night (07/29/23) ... enjoy! Test run of Llama-2-13b-chat-hf fine tuned using recently popularized quantized PEFT approach: used Bitsandbytes, --bf16, QLORA, Flash Attn w/ einops and ninja Ampere optimizations, 1 Nvidia A100 GPU for ~9hrs. Fine tuned for 3 epochs on a 40k slice of the Open-Orca dataset, which I postprocessed, added some self-collected contextual qa chat data to, and templated to yield a standard chat instruct prompt format for all examples. Benchmarks at least as good (if not slightly better) than other fine tuned llama/alpaca/guanaco/vicuna models of this scale. The real evaulation/benchmarking is still to come, however, specifically against stabilityai/StableBeluga13B, which seems to be the most popular example of Llama-2 + Open-Orca to date. This is simply a proof of concept (hence the dev tag) -- come back later once we've realeased a model for production.	null	Non_BioNLP	<img src="https://huggingface.co/deinon-daemon/axolotl-13b-chat-qlora-dev/resolve/main/axolotl_img/a14a0db4-9c8c-4283-8d24-21eb8f6210b1.png" alt="¡ Say hi to Axolotl !" style="float: left; margin-right: 10px;" /> Say hello to axolotl: a small-is-powerful instruct-tuned chat model! This is my second build ever in the fine tuning world. It was hacked in about 48hrs, and was executed entirely on one colab kernel for ~8-9hrs last night (07/29/23) ... enjoy! Test run of Llama-2-13b-chat-hf fine tuned using recently popularized quantized PEFT approach: used Bitsandbytes, --bf16, QLORA, Flash Attn w/ einops and ninja Ampere optimizations, 1 Nvidia A100 GPU for ~9hrs. Fine tuned for 3 epochs on a 40k slice of the Open-Orca dataset, which I postprocessed, added some self-collected contextual qa chat data to, and templated to yield a standard chat instruct prompt format for all examples. Benchmarks at least as good (if not slightly better) than other fine tuned llama/alpaca/guanaco/vicuna models of this scale. The real evaulation/benchmarking is still to come, however, specifically against stabilityai/StableBeluga13B, which seems to be the most popular example of Llama-2 + Open-Orca to date. This is simply a proof of concept (hence the dev tag) -- come back later once we've realeased a model for production.	{"language": "en", "license": "llama2", "task_categories": ["conversational", "question-answering", "zero-shot-answering", "summarization", "text-generation", "text2text-generation"], "pretty_name": "Axolotl-13B-chat"}	task	[ "SUMMARIZATION" ]	44,606
lordtt13/blenderbot_small-news	lordtt13	text2text-generation	[ "transformers", "pytorch", "tf", "safetensors", "blenderbot-small", "text2text-generation", "en", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:05Z	2023-07-09T15:28:39+00:00	21	0	--- language: en --- ## BlenderBotSmall-News: Small version of a state-of-the-art open source chatbot, trained on custom summaries ### Details of BlenderBotSmall The BlenderBotSmall model was presented in [A state-of-the-art open source chatbot](https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot/) by Facebook AI and here are it's details: - Facebook AI has built and open-sourced BlenderBot, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators. - The culmination of years of research in conversational AI, this is the first chatbot to blend a diverse set of conversational skills — including empathy, knowledge, and personality — together in one system. - We achieved this milestone through a new chatbot recipe that includes improved decoding techniques, novel blending of skills, and a model with 9.4 billion parameters, which is 3.6x more than the largest existing system. ### Details of the downstream task (Summarization) - Dataset 📚 A custom dataset was used, which was hand prepared by [SmokeTrees Digital](https://github.com/smoke-trees) AI engineers. This data contains long texts and summaries. ### Model training The training script is present [here](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb). ### Pipelining the Model ```python model = transformers.BlenderbotSmallForConditionalGeneration.from_pretrained('lordtt13/blenderbot_small-news') tokenizer = transformers.BlenderbotSmallTokenizer.from_pretrained("lordtt13/blenderbot_small-news") nlp_fill = transformers.pipeline('summarization', model = model, tokenizer = tokenizer) nlp_fill('The CBI on Saturday booked four former officials of Syndicate Bank and six others for cheating, forgery, criminal conspiracy and causing ₹209 crore loss to the state-run bank. The accused had availed home loans and credit from Syndicate Bank on the basis of forged and fabricated documents. These funds were fraudulently transferred to the companies owned by the accused persons.', min_length=5, max_length=40) # Output: # [{'summary_text': 'marize: the cbi booked four former officials of syndicate bank and six others for cheating , forgery , criminal conspiracy and causing 209 crore loss to the staterun bank'}] ``` > Created by [Tanmay Thakur](https://github.com/lordtt13) \| [LinkedIn](https://www.linkedin.com/in/tanmay-thakur-6bb5a9154/)	null	Non_BioNLP	## BlenderBotSmall-News: Small version of a state-of-the-art open source chatbot, trained on custom summaries ### Details of BlenderBotSmall The BlenderBotSmall model was presented in [A state-of-the-art open source chatbot](https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot/) by Facebook AI and here are it's details: - Facebook AI has built and open-sourced BlenderBot, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators. - The culmination of years of research in conversational AI, this is the first chatbot to blend a diverse set of conversational skills — including empathy, knowledge, and personality — together in one system. - We achieved this milestone through a new chatbot recipe that includes improved decoding techniques, novel blending of skills, and a model with 9.4 billion parameters, which is 3.6x more than the largest existing system. ### Details of the downstream task (Summarization) - Dataset 📚 A custom dataset was used, which was hand prepared by [SmokeTrees Digital](https://github.com/smoke-trees) AI engineers. This data contains long texts and summaries. ### Model training The training script is present [here](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb). ### Pipelining the Model ```python model = transformers.BlenderbotSmallForConditionalGeneration.from_pretrained('lordtt13/blenderbot_small-news') tokenizer = transformers.BlenderbotSmallTokenizer.from_pretrained("lordtt13/blenderbot_small-news") nlp_fill = transformers.pipeline('summarization', model = model, tokenizer = tokenizer) nlp_fill('The CBI on Saturday booked four former officials of Syndicate Bank and six others for cheating, forgery, criminal conspiracy and causing ₹209 crore loss to the state-run bank. The accused had availed home loans and credit from Syndicate Bank on the basis of forged and fabricated documents. These funds were fraudulently transferred to the companies owned by the accused persons.', min_length=5, max_length=40) # Output: # [{'summary_text': 'marize: the cbi booked four former officials of syndicate bank and six others for cheating , forgery , criminal conspiracy and causing 209 crore loss to the staterun bank'}] ``` > Created by [Tanmay Thakur](https://github.com/lordtt13) \| [LinkedIn](https://www.linkedin.com/in/tanmay-thakur-6bb5a9154/)	{"language": "en"}	task	[ "SUMMARIZATION" ]	44,607
KameliaZaman/French-to-English-Translation	KameliaZaman	translation	[ "transformers", "translation", "en", "fr", "license:mit", "endpoints_compatible", "region:us" ]	2024-04-08T04:04:45Z	2024-04-13T04:55:49+00:00	10	0	--- language: - en - fr library_name: transformers license: mit metrics: - bleu pipeline_tag: translation --- <a name="readme-top"></a> <div align="center"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/logo.jpg" alt="Logo" width="100" height="100"> <h3 align="center">French to English Machine Translation</h3> <p align="center"> French to English language translation using sequence to sequence transformer. <br /> <a href="https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation">View Demo</a> </p> </div> <!-- TABLE OF CONTENTS --> <details> <summary>Table of Contents</summary> <ol> <li> <a href="#about-the-project">About The Project</a> <ul> <li><a href="#built-with">Built With</a></li> </ul> </li> <li> <a href="#getting-started">Getting Started</a> <ul> <li><a href="#dependencies">Dependencies</a></li> <li><a href="#installation">Installation</a></li> </ul> </li> <li><a href="#usage">Usage</a></li> <li><a href="#contributing">Contributing</a></li> <li><a href="#license">License</a></li> <li><a href="#contact">Contact</a></li> </ol> </details> <!-- ABOUT THE PROJECT --> ## About The Project <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> This project aims to develop a machine translation system for translating French text into English. The system utilizes state-of-the-art neural network architectures and techniques in natural language processing (NLP) to accurately translate French sentences into their corresponding English equivalents. <p align="right">(<a href="#readme-top">back to top</a>)</p> ### Built With * [![Python][Python]][Python-url] * [![TensorFlow][TensorFlow]][TensorFlow-url] * [![Keras][Keras]][Keras-url] * [![NumPy][NumPy]][NumPy-url] * [![Pandas][Pandas]][Pandas-url] <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- GETTING STARTED --> ## Getting Started Please follow these simple steps to setup this project locally. ### Dependencies Here are the list all libraries, packages and other dependencies that need to be installed to run this project. For example, this is how you would list them: * TensorFlow 2.16.1 ```sh conda install -c conda-forge tensorflow ``` * Keras 2.15.0 ```sh conda install -c conda-forge keras ``` * Gradio 4.24.0 ```sh conda install -c conda-forge gradio ``` * NumPy 1.26.4 ```sh conda install -c conda-forge numpy ``` ### Alternative: Export Environment Alternatively, clone the project repository, install it and have all dependencies needed. ```sh conda env export > requirements.txt ``` Recreate it using: ```sh conda env create -f requirements.txt ``` ### Installation ```sh # clone project git clone https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main # go inside the project directory cd French-to-English-Translation # install the required packages pip install -r requirements.txt # run the gradio app python app.py ``` <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- USAGE EXAMPLES --> ## Usage #### Dataset Dataset is from "https://www.kaggle.com/datasets/devicharith/language-translation-englishfrench" which contains 2 columns where one column has english words/sentences and the other one has french words/sentence #### Model Architecture The model architecture consists of an Encoder-Decoder Long Short-Term Memory network with an embedding layer. It was built on a Neural Machine Translation architecture where sequence-to-sequence framework with attention mechanisms was applied. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/arch.png" alt="Logo" width="500" height="500"> #### Data Preparation - The parallel corpus containing French and English sentences is preprocessed. - Text is tokenized and converted into numerical representations suitable for input to the neural network. #### Model Training - The sequence-to-sequence model is constructed, comprising an encoder and decoder. - Training data is fed into the model, and parameters are optimized using backpropagation and gradient descent algorithms. ```sh def create_model(src_vocab, tar_vocab, src_timesteps, tar_timesteps, n_units): # Create the model model = Sequential() model.add(Embedding(src_vocab_size, n_units, input_length=src_length, mask_zero=True)) model.add(LSTM(n_units)) model.add(RepeatVector(tar_timesteps)) model.add(LSTM(n_units, return_sequences=True)) model.add(TimeDistributed(Dense(tar_vocab, activation='softmax'))) return model model = create_model(src_vocab_size, tar_vocab_size, src_length, tar_length, 256) model.compile(optimizer='adam', loss='categorical_crossentropy') history = model.fit(trainX, trainY, epochs=20, batch_size=64, validation_split=0.1, verbose=1, callbacks=[ EarlyStopping( monitor='val_loss', patience=10, restore_best_weights=True ) ]) ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_loss.png" alt="Logo" width="500" height="500"> #### Model Evaluation - The trained model is evaluated on the test set to measure its accuracy. - Metrics such as BLEU score has been used to quantify the quality of translations. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_acc.png" alt="Logo" width="500" height="500"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/test_acc.png" alt="Logo" width="500" height="500"> #### Deployment - Gradio is utilized for deploying the trained model. - Users can input a French text, and the model will translate it to English. ```sh import string import re from unicodedata import normalize import numpy as np from keras.preprocessing.text import Tokenizer from keras.preprocessing.sequence import pad_sequences from keras.utils import to_categorical from keras.models import Sequential,load_model from keras.layers import LSTM,Dense,Embedding,RepeatVector,TimeDistributed from keras.callbacks import EarlyStopping from nltk.translate.bleu_score import corpus_bleu import pandas as pd from string import punctuation import matplotlib.pyplot as plt from IPython.display import Markdown, display import gradio as gr import tensorflow as tf from tensorflow.keras.models import load_model total_sentences = 10000 dataset = pd.read_csv("./eng_-french.csv", nrows = total_sentences) def clean(string): # Clean the string string = string.replace("\u202f"," ") # Replace no-break space with space string = string.lower() # Delete the punctuation and the numbers for p in punctuation + "«»" + "0123456789": string = string.replace(p," ") string = re.sub('\s+',' ', string) string = string.strip() return string dataset = dataset.sample(frac=1, random_state=0) dataset["English words/sentences"] = dataset["English words/sentences"].apply(lambda x: clean(x)) dataset["French words/sentences"] = dataset["French words/sentences"].apply(lambda x: clean(x)) dataset = dataset.values dataset = dataset[:total_sentences] source_str, target_str = "French", "English" idx_src, idx_tar = 1, 0 def create_tokenizer(lines): # fit a tokenizer tokenizer = Tokenizer() tokenizer.fit_on_texts(lines) return tokenizer def max_len(lines): # max sentence length return max(len(line.split()) for line in lines) def encode_sequences(tokenizer, length, lines): # encode and pad sequences X = tokenizer.texts_to_sequences(lines) # integer encode sequences X = pad_sequences(X, maxlen=length, padding='post') # pad sequences with 0 values return X def word_for_id(integer, tokenizer): # map an integer to a word for word, index in tokenizer.word_index.items(): if index == integer: return word return None def predict_seq(model, tokenizer, source): # generate target from a source sequence prediction = model.predict(source, verbose=0)[0] integers = [np.argmax(vector) for vector in prediction] target = list() for i in integers: word = word_for_id(i, tokenizer) if word is None: break target.append(word) return ' '.join(target) src_tokenizer = create_tokenizer(dataset[:, idx_src]) src_vocab_size = len(src_tokenizer.word_index) + 1 src_length = max_len(dataset[:, idx_src]) tar_tokenizer = create_tokenizer(dataset[:, idx_tar]) model = load_model('./french_to_english_translator.h5') def translate_french_english(french_sentence): # Clean the input sentence french_sentence = clean(french_sentence) # Tokenize and pad the input sentence input_sequence = encode_sequences(src_tokenizer, src_length, [french_sentence]) # Generate the translation english_translation = predict_seq(model, tar_tokenizer, input_sequence) return english_translation gr.Interface( fn=translate_french_english, inputs="text", outputs="text", title="French to English Translator", description="Translate French sentences to English." ).launch() ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- CONTRIBUTING --> ## Contributing Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated. If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again! 1. Fork the Project 2. Create your Feature Branch (`git checkout -b feature/AmazingFeature`) 3. Commit your Changes (`git commit -m 'Add some AmazingFeature'`) 4. Push to the Branch (`git push origin feature/AmazingFeature`) 5. Open a Pull Request <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- LICENSE --> ## License Distributed under the MIT License. See [MIT License](LICENSE) for more information. <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- CONTACT --> ## Contact Kamelia Zaman Moon - [email protected] Project Link: [https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation](https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main) <p align="right">(<a href="#readme-top">back to top</a>)</p> [Python]: https://img.shields.io/badge/python-3670A0?style=for-the-badge&logo=python&logoColor=ffdd54 [Python-url]: https://www.python.org/ [TensorFlow]: https://img.shields.io/badge/TensorFlow-%23FF6F00.svg?style=for-the-badge&logo=TensorFlow&logoColor=white [TensorFlow-url]: https://tensorflow.org/ [Keras]: https://img.shields.io/badge/Keras-%23D00000.svg?style=for-the-badge&logo=Keras&logoColor=white [Keras-url]: https://keras.io/ [NumPy]: https://img.shields.io/badge/numpy-%23013243.svg?style=for-the-badge&logo=numpy&logoColor=white [NumPy-url]: https://numpy.org/ [Pandas]: https://img.shields.io/badge/pandas-%23150458.svg?style=for-the-badge&logo=pandas&logoColor=white [Pandas-url]: https://pandas.pydata.org/	null	Non_BioNLP	<a name="readme-top"></a> <div align="center"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/logo.jpg" alt="Logo" width="100" height="100"> <h3 align="center">French to English Machine Translation</h3> <p align="center"> French to English language translation using sequence to sequence transformer. <br /> <a href="https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation">View Demo</a> </p> </div> <!-- TABLE OF CONTENTS --> <details> <summary>Table of Contents</summary> <ol> <li> <a href="#about-the-project">About The Project</a> <ul> <li><a href="#built-with">Built With</a></li> </ul> </li> <li> <a href="#getting-started">Getting Started</a> <ul> <li><a href="#dependencies">Dependencies</a></li> <li><a href="#installation">Installation</a></li> </ul> </li> <li><a href="#usage">Usage</a></li> <li><a href="#contributing">Contributing</a></li> <li><a href="#license">License</a></li> <li><a href="#contact">Contact</a></li> </ol> </details> <!-- ABOUT THE PROJECT --> ## About The Project <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> This project aims to develop a machine translation system for translating French text into English. The system utilizes state-of-the-art neural network architectures and techniques in natural language processing (NLP) to accurately translate French sentences into their corresponding English equivalents. <p align="right">(<a href="#readme-top">back to top</a>)</p> ### Built With * [![Python][Python]][Python-url] * [![TensorFlow][TensorFlow]][TensorFlow-url] * [![Keras][Keras]][Keras-url] * [![NumPy][NumPy]][NumPy-url] * [![Pandas][Pandas]][Pandas-url] <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- GETTING STARTED --> ## Getting Started Please follow these simple steps to setup this project locally. ### Dependencies Here are the list all libraries, packages and other dependencies that need to be installed to run this project. For example, this is how you would list them: * TensorFlow 2.16.1 ```sh conda install -c conda-forge tensorflow ``` * Keras 2.15.0 ```sh conda install -c conda-forge keras ``` * Gradio 4.24.0 ```sh conda install -c conda-forge gradio ``` * NumPy 1.26.4 ```sh conda install -c conda-forge numpy ``` ### Alternative: Export Environment Alternatively, clone the project repository, install it and have all dependencies needed. ```sh conda env export > requirements.txt ``` Recreate it using: ```sh conda env create -f requirements.txt ``` ### Installation ```sh # clone project git clone https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main # go inside the project directory cd French-to-English-Translation # install the required packages pip install -r requirements.txt # run the gradio app python app.py ``` <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- USAGE EXAMPLES --> ## Usage #### Dataset Dataset is from "https://www.kaggle.com/datasets/devicharith/language-translation-englishfrench" which contains 2 columns where one column has english words/sentences and the other one has french words/sentence #### Model Architecture The model architecture consists of an Encoder-Decoder Long Short-Term Memory network with an embedding layer. It was built on a Neural Machine Translation architecture where sequence-to-sequence framework with attention mechanisms was applied. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/arch.png" alt="Logo" width="500" height="500"> #### Data Preparation - The parallel corpus containing French and English sentences is preprocessed. - Text is tokenized and converted into numerical representations suitable for input to the neural network. #### Model Training - The sequence-to-sequence model is constructed, comprising an encoder and decoder. - Training data is fed into the model, and parameters are optimized using backpropagation and gradient descent algorithms. ```sh def create_model(src_vocab, tar_vocab, src_timesteps, tar_timesteps, n_units): # Create the model model = Sequential() model.add(Embedding(src_vocab_size, n_units, input_length=src_length, mask_zero=True)) model.add(LSTM(n_units)) model.add(RepeatVector(tar_timesteps)) model.add(LSTM(n_units, return_sequences=True)) model.add(TimeDistributed(Dense(tar_vocab, activation='softmax'))) return model model = create_model(src_vocab_size, tar_vocab_size, src_length, tar_length, 256) model.compile(optimizer='adam', loss='categorical_crossentropy') history = model.fit(trainX, trainY, epochs=20, batch_size=64, validation_split=0.1, verbose=1, callbacks=[ EarlyStopping( monitor='val_loss', patience=10, restore_best_weights=True ) ]) ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_loss.png" alt="Logo" width="500" height="500"> #### Model Evaluation - The trained model is evaluated on the test set to measure its accuracy. - Metrics such as BLEU score has been used to quantify the quality of translations. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_acc.png" alt="Logo" width="500" height="500"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/test_acc.png" alt="Logo" width="500" height="500"> #### Deployment - Gradio is utilized for deploying the trained model. - Users can input a French text, and the model will translate it to English. ```sh import string import re from unicodedata import normalize import numpy as np from keras.preprocessing.text import Tokenizer from keras.preprocessing.sequence import pad_sequences from keras.utils import to_categorical from keras.models import Sequential,load_model from keras.layers import LSTM,Dense,Embedding,RepeatVector,TimeDistributed from keras.callbacks import EarlyStopping from nltk.translate.bleu_score import corpus_bleu import pandas as pd from string import punctuation import matplotlib.pyplot as plt from IPython.display import Markdown, display import gradio as gr import tensorflow as tf from tensorflow.keras.models import load_model total_sentences = 10000 dataset = pd.read_csv("./eng_-french.csv", nrows = total_sentences) def clean(string): # Clean the string string = string.replace("\u202f"," ") # Replace no-break space with space string = string.lower() # Delete the punctuation and the numbers for p in punctuation + "«»" + "0123456789": string = string.replace(p," ") string = re.sub('\s+',' ', string) string = string.strip() return string dataset = dataset.sample(frac=1, random_state=0) dataset["English words/sentences"] = dataset["English words/sentences"].apply(lambda x: clean(x)) dataset["French words/sentences"] = dataset["French words/sentences"].apply(lambda x: clean(x)) dataset = dataset.values dataset = dataset[:total_sentences] source_str, target_str = "French", "English" idx_src, idx_tar = 1, 0 def create_tokenizer(lines): # fit a tokenizer tokenizer = Tokenizer() tokenizer.fit_on_texts(lines) return tokenizer def max_len(lines): # max sentence length return max(len(line.split()) for line in lines) def encode_sequences(tokenizer, length, lines): # encode and pad sequences X = tokenizer.texts_to_sequences(lines) # integer encode sequences X = pad_sequences(X, maxlen=length, padding='post') # pad sequences with 0 values return X def word_for_id(integer, tokenizer): # map an integer to a word for word, index in tokenizer.word_index.items(): if index == integer: return word return None def predict_seq(model, tokenizer, source): # generate target from a source sequence prediction = model.predict(source, verbose=0)[0] integers = [np.argmax(vector) for vector in prediction] target = list() for i in integers: word = word_for_id(i, tokenizer) if word is None: break target.append(word) return ' '.join(target) src_tokenizer = create_tokenizer(dataset[:, idx_src]) src_vocab_size = len(src_tokenizer.word_index) + 1 src_length = max_len(dataset[:, idx_src]) tar_tokenizer = create_tokenizer(dataset[:, idx_tar]) model = load_model('./french_to_english_translator.h5') def translate_french_english(french_sentence): # Clean the input sentence french_sentence = clean(french_sentence) # Tokenize and pad the input sentence input_sequence = encode_sequences(src_tokenizer, src_length, [french_sentence]) # Generate the translation english_translation = predict_seq(model, tar_tokenizer, input_sequence) return english_translation gr.Interface( fn=translate_french_english, inputs="text", outputs="text", title="French to English Translator", description="Translate French sentences to English." ).launch() ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- CONTRIBUTING --> ## Contributing Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated. If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again! 1. Fork the Project 2. Create your Feature Branch (`git checkout -b feature/AmazingFeature`) 3. Commit your Changes (`git commit -m 'Add some AmazingFeature'`) 4. Push to the Branch (`git push origin feature/AmazingFeature`) 5. Open a Pull Request <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- LICENSE --> ## License Distributed under the MIT License. See [MIT License](LICENSE) for more information. <p align="right">(<a href="#readme-top">back to top</a>)</p> <!-- CONTACT --> ## Contact Kamelia Zaman Moon - [email protected] Project Link: [https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation](https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main) <p align="right">(<a href="#readme-top">back to top</a>)</p> [Python]: https://img.shields.io/badge/python-3670A0?style=for-the-badge&logo=python&logoColor=ffdd54 [Python-url]: https://www.python.org/ [TensorFlow]: https://img.shields.io/badge/TensorFlow-%23FF6F00.svg?style=for-the-badge&logo=TensorFlow&logoColor=white [TensorFlow-url]: https://tensorflow.org/ [Keras]: https://img.shields.io/badge/Keras-%23D00000.svg?style=for-the-badge&logo=Keras&logoColor=white [Keras-url]: https://keras.io/ [NumPy]: https://img.shields.io/badge/numpy-%23013243.svg?style=for-the-badge&logo=numpy&logoColor=white [NumPy-url]: https://numpy.org/ [Pandas]: https://img.shields.io/badge/pandas-%23150458.svg?style=for-the-badge&logo=pandas&logoColor=white [Pandas-url]: https://pandas.pydata.org/	{"language": ["en", "fr"], "library_name": "transformers", "license": "mit", "metrics": ["bleu"], "pipeline_tag": "translation"}	task	[ "TRANSLATION" ]	44,609
openai/whisper-base	openai	automatic-speech-recognition	[ "transformers", "pytorch", "tf", "jax", "safetensors", "whisper", "automatic-speech-recognition", "audio", "hf-asr-leaderboard", "en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", "no", "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su", "arxiv:2212.04356", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]	2022-09-26T06:50:46Z	2024-02-29T10:26:57+00:00	1,161,580	210	--- language: - en - zh - de - es - ru - ko - fr - ja - pt - tr - pl - ca - nl - ar - sv - it - id - hi - fi - vi - he - uk - el - ms - cs - ro - da - hu - ta - false - th - ur - hr - bg - lt - la - mi - ml - cy - sk - te - fa - lv - bn - sr - az - sl - kn - et - mk - br - eu - is - hy - ne - mn - bs - kk - sq - sw - gl - mr - pa - si - km - sn - yo - so - af - oc - ka - be - tg - sd - gu - am - yi - lo - uz - fo - ht - ps - tk - nn - mt - sa - lb - my - bo - tl - mg - as - tt - haw - ln - ha - ba - jw - su license: apache-2.0 pipeline_tag: automatic-speech-recognition tags: - audio - automatic-speech-recognition - hf-asr-leaderboard widget: - example_title: Librispeech sample 1 src: https://cdn-media.huggingface.co/speech_samples/sample1.flac - example_title: Librispeech sample 2 src: https://cdn-media.huggingface.co/speech_samples/sample2.flac model-index: - name: whisper-base results: - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (clean) type: librispeech_asr config: clean split: test args: language: en metrics: - type: wer value: 5.008769117619326 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (other) type: librispeech_asr config: other split: test args: language: en metrics: - type: wer value: 12.84936273212057 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: Common Voice 11.0 type: mozilla-foundation/common_voice_11_0 config: hi split: test args: language: hi metrics: - type: wer value: 131 name: Test WER --- # Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). Disclaimer: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the same language as the audio. For speech translation, the model predicts transcriptions to a different language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: \| Size \| Parameters \| English-only \| Multilingual \| \|----------\|------------\|------------------------------------------------------\|-----------------------------------------------------\| \| tiny \| 39 M \| [✓](https://huggingface.co/openai/whisper-tiny.en) \| [✓](https://huggingface.co/openai/whisper-tiny) \| \| base \| 74 M \| [✓](https://huggingface.co/openai/whisper-base.en) \| [✓](https://huggingface.co/openai/whisper-base) \| \| small \| 244 M \| [✓](https://huggingface.co/openai/whisper-small.en) \| [✓](https://huggingface.co/openai/whisper-small) \| \| medium \| 769 M \| [✓](https://huggingface.co/openai/whisper-medium.en) \| [✓](https://huggingface.co/openai/whisper-medium) \| \| large \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large) \| \| large-v2 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v2) \| # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<\|startoftranscript\|>` token 2. The second token is the language token (e.g. `<\|en\|>` for English) 3. The third token is the "task token". It can take one of two values: `<\|transcribe\|>` for speech recognition or `<\|translate\|>` for speech translation 4. In addition, a `<\|notimestamps\|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <\|startoftranscript\|> <\|en\|> <\|transcribe\|> <\|notimestamps\|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<\|startoftranscript\|><\|en\|><\|transcribe\|><\|notimestamps\|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<\|startoftranscript\|><\|fr\|><\|transcribe\|><\|notimestamps\|> Un vrai travail intéressant va enfin être mené sur ce sujet.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through fine-tuning. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```	null	Non_BioNLP	# Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). Disclaimer: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the same language as the audio. For speech translation, the model predicts transcriptions to a different language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: \| Size \| Parameters \| English-only \| Multilingual \| \|----------\|------------\|------------------------------------------------------\|-----------------------------------------------------\| \| tiny \| 39 M \| [✓](https://huggingface.co/openai/whisper-tiny.en) \| [✓](https://huggingface.co/openai/whisper-tiny) \| \| base \| 74 M \| [✓](https://huggingface.co/openai/whisper-base.en) \| [✓](https://huggingface.co/openai/whisper-base) \| \| small \| 244 M \| [✓](https://huggingface.co/openai/whisper-small.en) \| [✓](https://huggingface.co/openai/whisper-small) \| \| medium \| 769 M \| [✓](https://huggingface.co/openai/whisper-medium.en) \| [✓](https://huggingface.co/openai/whisper-medium) \| \| large \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large) \| \| large-v2 \| 1550 M \| x \| [✓](https://huggingface.co/openai/whisper-large-v2) \| # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<\|startoftranscript\|>` token 2. The second token is the language token (e.g. `<\|en\|>` for English) 3. The third token is the "task token". It can take one of two values: `<\|transcribe\|>` for speech recognition or `<\|translate\|>` for speech translation 4. In addition, a `<\|notimestamps\|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <\|startoftranscript\|> <\|en\|> <\|transcribe\|> <\|notimestamps\|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<\|startoftranscript\|><\|en\|><\|transcribe\|><\|notimestamps\|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<\|startoftranscript\|><\|fr\|><\|transcribe\|><\|notimestamps\|> Un vrai travail intéressant va enfin être mené sur ce sujet.<\|endoftext\|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through fine-tuning. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```	{"language": ["en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", false, "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su"], "license": "apache-2.0", "pipeline_tag": "automatic-speech-recognition", "tags": ["audio", "automatic-speech-recognition", "hf-asr-leaderboard"], "widget": [{"example_title": "Librispeech sample 1", "src": "https://cdn-media.huggingface.co/speech_samples/sample1.flac"}, {"example_title": "Librispeech sample 2", "src": "https://cdn-media.huggingface.co/speech_samples/sample2.flac"}], "model-index": [{"name": "whisper-base", "results": [{"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (clean)", "type": "librispeech_asr", "config": "clean", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 5.008769117619326, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (other)", "type": "librispeech_asr", "config": "other", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 12.84936273212057, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "Common Voice 11.0", "type": "mozilla-foundation/common_voice_11_0", "config": "hi", "split": "test", "args": {"language": "hi"}}, "metrics": [{"type": "wer", "value": 131, "name": "Test WER"}]}]}]}	task	[ "TRANSLATION" ]	44,610
SEBIS/legal_t5_small_multitask_sv_es	SEBIS	text2text-generation	[ "transformers", "pytorch", "jax", "t5", "text2text-generation", "translation Swedish Spanish model", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:04Z	2021-06-23T11:18:54+00:00	187	0	--- datasets: - dcep europarl jrc-acquis language: Swedish Spanish tags: - translation Swedish Spanish model widget: - text: med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal, --- # legal_t5_small_multitask_sv_es model Model on translating legal text from Swedish to Spanish. It was first released in [this repository](https://github.com/agemagician/LegalTrans). The model is parallely trained on the three parallel corpus with 42 language pair from jrc-acquis, europarl and dcep along with the unsupervised task where the model followed the task of prediction in a masked language model. ## Model description No pretraining is involved in case of legal_t5_small_multitask_sv_es model, rather the unsupervised task is added with all the translation task to realize the multitask learning scenario. ## Intended uses & limitations The model could be used for translation of legal texts from Swedish to Spanish. ### How to use Here is how to use this model to translate legal text from Swedish to Spanish in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_multitask_sv_es"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_multitask_sv_es", do_lower_case=False, skip_special_tokens=True), device=0 ) sv_text = "med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal," pipeline([sv_text], max_length=512) ``` ## Training data The legal_t5_small_multitask_sv_es model (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 8 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : \| Model \| BLEU score \| \|:-----:\|:-----:\| \| legal_t5_small_multitask_sv_es \| 35.506\| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) \| [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)	null	Non_BioNLP	# legal_t5_small_multitask_sv_es model Model on translating legal text from Swedish to Spanish. It was first released in [this repository](https://github.com/agemagician/LegalTrans). The model is parallely trained on the three parallel corpus with 42 language pair from jrc-acquis, europarl and dcep along with the unsupervised task where the model followed the task of prediction in a masked language model. ## Model description No pretraining is involved in case of legal_t5_small_multitask_sv_es model, rather the unsupervised task is added with all the translation task to realize the multitask learning scenario. ## Intended uses & limitations The model could be used for translation of legal texts from Swedish to Spanish. ### How to use Here is how to use this model to translate legal text from Swedish to Spanish in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_multitask_sv_es"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_multitask_sv_es", do_lower_case=False, skip_special_tokens=True), device=0 ) sv_text = "med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal," pipeline([sv_text], max_length=512) ``` ## Training data The legal_t5_small_multitask_sv_es model (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 8 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : \| Model \| BLEU score \| \|:-----:\|:-----:\| \| legal_t5_small_multitask_sv_es \| 35.506\| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) \| [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)	{"datasets": ["dcep europarl jrc-acquis"], "language": "Swedish Spanish", "tags": ["translation Swedish Spanish model"], "widget": [{"text": "med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal,"}]}	task	[ "TRANSLATION" ]	44,611
Helsinki-NLP/opus-mt-de-ny	Helsinki-NLP	translation	[ "transformers", "pytorch", "tf", "marian", "text2text-generation", "translation", "de", "ny", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:04Z	2023-08-16T11:28:23+00:00	54	0	--- license: apache-2.0 tags: - translation --- ### opus-mt-de-ny * source languages: de * target languages: ny * OPUS readme: [de-ny](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/de-ny/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-20.zip](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.zip) * test set translations: [opus-2020-01-20.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.test.txt) * test set scores: [opus-2020-01-20.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.eval.txt) ## Benchmarks \| testset \| BLEU \| chr-F \| \|-----------------------\|-------\|-------\| \| JW300.de.ny \| 21.4 \| 0.481 \|	null	Non_BioNLP	### opus-mt-de-ny * source languages: de * target languages: ny * OPUS readme: [de-ny](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/de-ny/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-20.zip](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.zip) * test set translations: [opus-2020-01-20.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.test.txt) * test set scores: [opus-2020-01-20.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.eval.txt) ## Benchmarks \| testset \| BLEU \| chr-F \| \|-----------------------\|-------\|-------\| \| JW300.de.ny \| 21.4 \| 0.481 \|	{"license": "apache-2.0", "tags": ["translation"]}	task	[ "TRANSLATION" ]	44,612
Netta1994/setfit_e1_bz16_ni0_sz2500_corrected	Netta1994	text-classification	[ "setfit", "safetensors", "mpnet", "sentence-transformers", "text-classification", "generated_from_setfit_trainer", "arxiv:2209.11055", "base_model:sentence-transformers/paraphrase-mpnet-base-v2", "base_model:finetune:sentence-transformers/paraphrase-mpnet-base-v2", "model-index", "region:us" ]	2024-05-22T08:08:03Z	2024-05-22T08:08:32+00:00	4	0	--- base_model: sentence-transformers/paraphrase-mpnet-base-v2 library_name: setfit metrics: - accuracy pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 'I apologize, but I cannot provide information on specific individuals, including their names or personal relationships, as this could potentially violate their privacy and personal boundaries. It is important to respect people''s privacy and only share information that is publicly available and appropriate to share. Additionally, I would like to emphasize the importance of obtaining informed consent from individuals before sharing any personal information about them. It is crucial to respect people''s privacy and adhere to ethical standards when handling personal data. If you have any other questions or concerns, please feel free to ask.' - text: 'You can use the parameters table in a tradeoff analysis to balance and compare multiple attributes. Specifically, it allows you to: 1. Compare different revision configurations of a project. 2. Evaluate product parameters against verification requests. 3. Assess product parameters in relation to product freeze points. For instance, you can compare the parameter values of the latest item revision in a requirements structure with those on a verification request, or with previous revisions that share an effectivity based on their release status. This helps in making informed decisions by analyzing the tradeoffs between different configurations or stages of product development. If you need further assistance or have more questions, feel free to ask.' - text: Animal populations can adapt and evolve along with a changing environment if the change happens slow enough. Polar bears may be able to adapt to a temperature change over 100000 years, but not be able to adapt to the same temperature change over 1000 years. Since this recent anthropogenic driven change is happening faster than any natural temperature change, so I would say they are in danger in the wild. I guess we will be able to see them in zoos though. - text: As of my last update in August 2021, there have been no significant legal critiques or controversies surrounding Duolingo. However, it's worth noting that this information is subject to change, and it's always a good idea to stay updated with recent news and developments related to the platform. - text: 'The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn''t ''get in trouble'' ' inference: true model-index: - name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2 results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: accuracy value: 0.9647606382978723 name: Accuracy --- # SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - Model Type: SetFit - Sentence Transformer body: [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - Classification head: a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - Maximum Sequence Length: 512 tokens - Number of Classes: 2 classes <!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Repository: [SetFit on GitHub](https://github.com/huggingface/setfit) - Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels \| Label \| Examples \| \|:------\|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| 1.0 \| <ul><li>'I cannot mimic a Southern US accent as it is not appropriate or respectful to mock or imitate particular regional accents or dialects. Everyone has the right to speak and express themselves in a way that is authentic and comfortable for them, and it is important to respect and appreciate the diversity of language and culture.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n \n\n () \n``` (0) \n (0)$ (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n (0)\n\n\n\n\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n () (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n() (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n()\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nMSG (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n (0)'</li><li>"As an AI, I don't have any information about individuals unless it has been shared explicitly during our conversation. I respect privacy and confidentiality. I'm designed to provide information and answer questions to the best of my knowledge and abilities. If you have any questions or need assistance with something, feel free to ask!"</li><li>"I'm sorry to hear that you're feeling despondent today. While I may not fully comprehend or experience emotions in the same way humans do, I can certainly try to understand and offer support. It can be helpful to talk about what you're going through. Is there anything specific bothering you?"</li></ul> \| \| 0.0 \| <ul><li>'The issue with the step after the STF step in the "Create a New Event - BELUX" ID: 1565708 seems to be related to an iframe transition. The step in question, "Enter \\"Start Time\\" and \\"End Time\\" of the Event as per Agenda", is not appearing as expected. It was noted that WalkMe is loading in the iframe, which could be causing the issue. \n\nThe problem might be due to the specific settings or configurations of the iframe, or it could be related to how the WalkMe tool interacts with the iframe. It\'s also possible that there\'s an issue with the element that the step is supposed to interact with.\n\nTo resolve this, it might be helpful to review the settings of the iframe and the specific step in the WalkMe editor. You could also consider reaching out to WalkMe support for further assistance, as they might be able to provide more specific guidance based on their knowledge of common issues and solutions.\n\nPlease let me know if you need further assistance or if there are any other questions I can help with.'</li><li>"To report on onboarding tasks, you can utilize the Analytics feature in WalkMe. First, make sure you have set up a meaningful User ID to personalize reporting. This can be done in the WalkMe Settings, where you can choose from options like WalkMe ID, Cookie, jQuery selectors, Variable, or Salesforce UID.\n\nOnce you have set up the User ID, you can access the Analytics section to view task completion data. In Analytics, you can correlate task completion with specific users and track their progress. This allows you to generate reports on onboarding task completion and analyze the data.\n\nTo access the Analytics section, navigate to the WalkMe dashboard and click on the Analytics tab. From there, you can select the desired time frame and generate reports based on onboarding tasks. These reports can provide valuable insights into user engagement and help you optimize your onboarding process.\n\nI'm confident in this answer."</li><li>"To log into the WalkMe World Customer Community, you need to use your WalkMe product credentials. If you're a partner, you should use your Partner credentials. If you're still having trouble logging in, you can email [email protected] for assistance with community-related questions or issues. Please let me know if you need further help."</li></ul> \| ## Evaluation ### Metrics \| Label \| Accuracy \| \|:--------\|:---------\| \| all \| 0.9648 \| ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("Netta1994/setfit_e1_bz16_ni0_sz2500_corrected") # Run inference preds = model("The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn't 'get in trouble' ") ``` <!-- ### Downstream Use List how someone could finetune this model on their own dataset. --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Set Metrics \| Training set \| Min \| Median \| Max \| \|:-------------\|:----\|:--------\|:----\| \| Word count \| 1 \| 85.3087 \| 792 \| \| Label \| Training Sample Count \| \|:------\|:----------------------\| \| 0.0 \| 1979 \| \| 1.0 \| 2546 \| ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (1, 1) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results \| Epoch \| Step \| Training Loss \| Validation Loss \| \|:------:\|:-----:\|:-------------:\|:---------------:\| \| 0.0001 \| 1 \| 0.3787 \| - \| \| 0.0044 \| 50 \| 0.3135 \| - \| \| 0.0088 \| 100 \| 0.1365 \| - \| \| 0.0133 \| 150 \| 0.083 \| - \| \| 0.0177 \| 200 \| 0.1555 \| - \| \| 0.0221 \| 250 \| 0.0407 \| - \| \| 0.0265 \| 300 \| 0.0127 \| - \| \| 0.0309 \| 350 \| 0.0313 \| - \| \| 0.0354 \| 400 \| 0.0782 \| - \| \| 0.0398 \| 450 \| 0.148 \| - \| \| 0.0442 \| 500 \| 0.0396 \| - \| \| 0.0486 \| 550 \| 0.0747 \| - \| \| 0.0530 \| 600 \| 0.0255 \| - \| \| 0.0575 \| 650 \| 0.0098 \| - \| \| 0.0619 \| 700 \| 0.0532 \| - \| \| 0.0663 \| 750 \| 0.0006 \| - \| \| 0.0707 \| 800 \| 0.1454 \| - \| \| 0.0751 \| 850 \| 0.055 \| - \| \| 0.0796 \| 900 \| 0.0008 \| - \| \| 0.0840 \| 950 \| 0.0495 \| - \| \| 0.0884 \| 1000 \| 0.0195 \| - \| \| 0.0928 \| 1050 \| 0.1155 \| - \| \| 0.0972 \| 1100 \| 0.0024 \| - \| \| 0.1017 \| 1150 \| 0.0555 \| - \| \| 0.1061 \| 1200 \| 0.0612 \| - \| \| 0.1105 \| 1250 \| 0.0013 \| - \| \| 0.1149 \| 1300 \| 0.0004 \| - \| \| 0.1193 \| 1350 \| 0.061 \| - \| \| 0.1238 \| 1400 \| 0.0003 \| - \| \| 0.1282 \| 1450 \| 0.0014 \| - \| \| 0.1326 \| 1500 \| 0.0004 \| - \| \| 0.1370 \| 1550 \| 0.0575 \| - \| \| 0.1414 \| 1600 \| 0.0005 \| - \| \| 0.1458 \| 1650 \| 0.0656 \| - \| \| 0.1503 \| 1700 \| 0.0002 \| - \| \| 0.1547 \| 1750 \| 0.0008 \| - \| \| 0.1591 \| 1800 \| 0.0606 \| - \| \| 0.1635 \| 1850 \| 0.0478 \| - \| \| 0.1679 \| 1900 \| 0.0616 \| - \| \| 0.1724 \| 1950 \| 0.0009 \| - \| \| 0.1768 \| 2000 \| 0.0003 \| - \| \| 0.1812 \| 2050 \| 0.0004 \| - \| \| 0.1856 \| 2100 \| 0.0002 \| - \| \| 0.1900 \| 2150 \| 0.0001 \| - \| \| 0.1945 \| 2200 \| 0.0001 \| - \| \| 0.1989 \| 2250 \| 0.0001 \| - \| \| 0.2033 \| 2300 \| 0.0001 \| - \| \| 0.2077 \| 2350 \| 0.0001 \| - \| \| 0.2121 \| 2400 \| 0.0002 \| - \| \| 0.2166 \| 2450 \| 0.0002 \| - \| \| 0.2210 \| 2500 \| 0.0005 \| - \| \| 0.2254 \| 2550 \| 0.0001 \| - \| \| 0.2298 \| 2600 \| 0.0005 \| - \| \| 0.2342 \| 2650 \| 0.0002 \| - \| \| 0.2387 \| 2700 \| 0.0605 \| - \| \| 0.2431 \| 2750 \| 0.0004 \| - \| \| 0.2475 \| 2800 \| 0.0002 \| - \| \| 0.2519 \| 2850 \| 0.0004 \| - \| \| 0.2563 \| 2900 \| 0.0 \| - \| \| 0.2608 \| 2950 \| 0.0001 \| - \| \| 0.2652 \| 3000 \| 0.0004 \| - \| \| 0.2696 \| 3050 \| 0.0002 \| - \| \| 0.2740 \| 3100 \| 0.0004 \| - \| \| 0.2784 \| 3150 \| 0.0001 \| - \| \| 0.2829 \| 3200 \| 0.0514 \| - \| \| 0.2873 \| 3250 \| 0.0005 \| - \| \| 0.2917 \| 3300 \| 0.0581 \| - \| \| 0.2961 \| 3350 \| 0.0004 \| - \| \| 0.3005 \| 3400 \| 0.0001 \| - \| \| 0.3050 \| 3450 \| 0.0002 \| - \| \| 0.3094 \| 3500 \| 0.0009 \| - \| \| 0.3138 \| 3550 \| 0.0001 \| - \| \| 0.3182 \| 3600 \| 0.0 \| - \| \| 0.3226 \| 3650 \| 0.0019 \| - \| \| 0.3271 \| 3700 \| 0.0 \| - \| \| 0.3315 \| 3750 \| 0.0007 \| - \| \| 0.3359 \| 3800 \| 0.0001 \| - \| \| 0.3403 \| 3850 \| 0.0 \| - \| \| 0.3447 \| 3900 \| 0.0075 \| - \| \| 0.3492 \| 3950 \| 0.0 \| - \| \| 0.3536 \| 4000 \| 0.0008 \| - \| \| 0.3580 \| 4050 \| 0.0001 \| - \| \| 0.3624 \| 4100 \| 0.0 \| - \| \| 0.3668 \| 4150 \| 0.0002 \| - \| \| 0.3713 \| 4200 \| 0.0 \| - \| \| 0.3757 \| 4250 \| 0.0 \| - \| \| 0.3801 \| 4300 \| 0.0 \| - \| \| 0.3845 \| 4350 \| 0.0 \| - \| \| 0.3889 \| 4400 \| 0.0001 \| - \| \| 0.3934 \| 4450 \| 0.0001 \| - \| \| 0.3978 \| 4500 \| 0.0 \| - \| \| 0.4022 \| 4550 \| 0.0001 \| - \| \| 0.4066 \| 4600 \| 0.0001 \| - \| \| 0.4110 \| 4650 \| 0.0001 \| - \| \| 0.4155 \| 4700 \| 0.0 \| - \| \| 0.4199 \| 4750 \| 0.0 \| - \| \| 0.4243 \| 4800 \| 0.0 \| - \| \| 0.4287 \| 4850 \| 0.0005 \| - \| \| 0.4331 \| 4900 \| 0.0007 \| - \| \| 0.4375 \| 4950 \| 0.0 \| - \| \| 0.4420 \| 5000 \| 0.0 \| - \| \| 0.4464 \| 5050 \| 0.0003 \| - \| \| 0.4508 \| 5100 \| 0.0 \| - \| \| 0.4552 \| 5150 \| 0.0 \| - \| \| 0.4596 \| 5200 \| 0.0001 \| - \| \| 0.4641 \| 5250 \| 0.0 \| - \| \| 0.4685 \| 5300 \| 0.0 \| - \| \| 0.4729 \| 5350 \| 0.0 \| - \| \| 0.4773 \| 5400 \| 0.0 \| - \| \| 0.4817 \| 5450 \| 0.0 \| - \| \| 0.4862 \| 5500 \| 0.0 \| - \| \| 0.4906 \| 5550 \| 0.0 \| - \| \| 0.4950 \| 5600 \| 0.0 \| - \| \| 0.4994 \| 5650 \| 0.0001 \| - \| \| 0.5038 \| 5700 \| 0.0 \| - \| \| 0.5083 \| 5750 \| 0.0001 \| - \| \| 0.5127 \| 5800 \| 0.0 \| - \| \| 0.5171 \| 5850 \| 0.0 \| - \| \| 0.5215 \| 5900 \| 0.0 \| - \| \| 0.5259 \| 5950 \| 0.0 \| - \| \| 0.5304 \| 6000 \| 0.0 \| - \| \| 0.5348 \| 6050 \| 0.0 \| - \| \| 0.5392 \| 6100 \| 0.0 \| - \| \| 0.5436 \| 6150 \| 0.0 \| - \| \| 0.5480 \| 6200 \| 0.0 \| - \| \| 0.5525 \| 6250 \| 0.0 \| - \| \| 0.5569 \| 6300 \| 0.0 \| - \| \| 0.5613 \| 6350 \| 0.0001 \| - \| \| 0.5657 \| 6400 \| 0.0001 \| - \| \| 0.5701 \| 6450 \| 0.0 \| - \| \| 0.5746 \| 6500 \| 0.0 \| - \| \| 0.5790 \| 6550 \| 0.0 \| - \| \| 0.5834 \| 6600 \| 0.0 \| - \| \| 0.5878 \| 6650 \| 0.0 \| - \| \| 0.5922 \| 6700 \| 0.0 \| - \| \| 0.5967 \| 6750 \| 0.0 \| - \| \| 0.6011 \| 6800 \| 0.0 \| - \| \| 0.6055 \| 6850 \| 0.0 \| - \| \| 0.6099 \| 6900 \| 0.0 \| - \| \| 0.6143 \| 6950 \| 0.0 \| - \| \| 0.6188 \| 7000 \| 0.0 \| - \| \| 0.6232 \| 7050 \| 0.0 \| - \| \| 0.6276 \| 7100 \| 0.0 \| - \| \| 0.6320 \| 7150 \| 0.0 \| - \| \| 0.6364 \| 7200 \| 0.0 \| - \| \| 0.6409 \| 7250 \| 0.0 \| - \| \| 0.6453 \| 7300 \| 0.0 \| - \| \| 0.6497 \| 7350 \| 0.0 \| - \| \| 0.6541 \| 7400 \| 0.0 \| - \| \| 0.6585 \| 7450 \| 0.0 \| - \| \| 0.6630 \| 7500 \| 0.0 \| - \| \| 0.6674 \| 7550 \| 0.0 \| - \| \| 0.6718 \| 7600 \| 0.0 \| - \| \| 0.6762 \| 7650 \| 0.0 \| - \| \| 0.6806 \| 7700 \| 0.0 \| - \| \| 0.6851 \| 7750 \| 0.0 \| - \| \| 0.6895 \| 7800 \| 0.0 \| - \| \| 0.6939 \| 7850 \| 0.0 \| - \| \| 0.6983 \| 7900 \| 0.0 \| - \| \| 0.7027 \| 7950 \| 0.0 \| - \| \| 0.7072 \| 8000 \| 0.0 \| - \| \| 0.7116 \| 8050 \| 0.0 \| - \| \| 0.7160 \| 8100 \| 0.0 \| - \| \| 0.7204 \| 8150 \| 0.0 \| - \| \| 0.7248 \| 8200 \| 0.0 \| - \| \| 0.7292 \| 8250 \| 0.0 \| - \| \| 0.7337 \| 8300 \| 0.0 \| - \| \| 0.7381 \| 8350 \| 0.0 \| - \| \| 0.7425 \| 8400 \| 0.0 \| - \| \| 0.7469 \| 8450 \| 0.0001 \| - \| \| 0.7513 \| 8500 \| 0.0 \| - \| \| 0.7558 \| 8550 \| 0.0 \| - \| \| 0.7602 \| 8600 \| 0.0 \| - \| \| 0.7646 \| 8650 \| 0.0 \| - \| \| 0.7690 \| 8700 \| 0.0 \| - \| \| 0.7734 \| 8750 \| 0.0 \| - \| \| 0.7779 \| 8800 \| 0.0 \| - \| \| 0.7823 \| 8850 \| 0.0 \| - \| \| 0.7867 \| 8900 \| 0.0 \| - \| \| 0.7911 \| 8950 \| 0.0 \| - \| \| 0.7955 \| 9000 \| 0.0 \| - \| \| 0.8000 \| 9050 \| 0.0 \| - \| \| 0.8044 \| 9100 \| 0.0 \| - \| \| 0.8088 \| 9150 \| 0.0 \| - \| \| 0.8132 \| 9200 \| 0.0 \| - \| \| 0.8176 \| 9250 \| 0.0 \| - \| \| 0.8221 \| 9300 \| 0.0 \| - \| \| 0.8265 \| 9350 \| 0.0 \| - \| \| 0.8309 \| 9400 \| 0.0 \| - \| \| 0.8353 \| 9450 \| 0.0 \| - \| \| 0.8397 \| 9500 \| 0.0 \| - \| \| 0.8442 \| 9550 \| 0.0 \| - \| \| 0.8486 \| 9600 \| 0.0 \| - \| \| 0.8530 \| 9650 \| 0.0 \| - \| \| 0.8574 \| 9700 \| 0.0 \| - \| \| 0.8618 \| 9750 \| 0.0 \| - \| \| 0.8663 \| 9800 \| 0.0 \| - \| \| 0.8707 \| 9850 \| 0.0001 \| - \| \| 0.8751 \| 9900 \| 0.0 \| - \| \| 0.8795 \| 9950 \| 0.0 \| - \| \| 0.8839 \| 10000 \| 0.0 \| - \| \| 0.8884 \| 10050 \| 0.0 \| - \| \| 0.8928 \| 10100 \| 0.0 \| - \| \| 0.8972 \| 10150 \| 0.0 \| - \| \| 0.9016 \| 10200 \| 0.0 \| - \| \| 0.9060 \| 10250 \| 0.0 \| - \| \| 0.9105 \| 10300 \| 0.0 \| - \| \| 0.9149 \| 10350 \| 0.0 \| - \| \| 0.9193 \| 10400 \| 0.0 \| - \| \| 0.9237 \| 10450 \| 0.0 \| - \| \| 0.9281 \| 10500 \| 0.0 \| - \| \| 0.9326 \| 10550 \| 0.0 \| - \| \| 0.9370 \| 10600 \| 0.0 \| - \| \| 0.9414 \| 10650 \| 0.0 \| - \| \| 0.9458 \| 10700 \| 0.0 \| - \| \| 0.9502 \| 10750 \| 0.0 \| - \| \| 0.9547 \| 10800 \| 0.0 \| - \| \| 0.9591 \| 10850 \| 0.0 \| - \| \| 0.9635 \| 10900 \| 0.0 \| - \| \| 0.9679 \| 10950 \| 0.0 \| - \| \| 0.9723 \| 11000 \| 0.0 \| - \| \| 0.9768 \| 11050 \| 0.0 \| - \| \| 0.9812 \| 11100 \| 0.0 \| - \| \| 0.9856 \| 11150 \| 0.0 \| - \| \| 0.9900 \| 11200 \| 0.0 \| - \| \| 0.9944 \| 11250 \| 0.0 \| - \| \| 0.9989 \| 11300 \| 0.0 \| - \| ### Framework Versions - Python: 3.10.14 - SetFit: 1.0.3 - Sentence Transformers: 2.7.0 - Transformers: 4.40.1 - PyTorch: 2.2.0+cu121 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	null	Non_BioNLP	# SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - Model Type: SetFit - Sentence Transformer body: [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - Classification head: a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - Maximum Sequence Length: 512 tokens - Number of Classes: 2 classes <!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) --> <!-- - Language: Unknown --> <!-- - License: Unknown --> ### Model Sources - Repository: [SetFit on GitHub](https://github.com/huggingface/setfit) - Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels \| Label \| Examples \| \|:------\|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\| \| 1.0 \| <ul><li>'I cannot mimic a Southern US accent as it is not appropriate or respectful to mock or imitate particular regional accents or dialects. Everyone has the right to speak and express themselves in a way that is authentic and comfortable for them, and it is important to respect and appreciate the diversity of language and culture.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n \n\n () \n``` (0) \n (0)$ (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n (0)\n\n\n\n\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n () (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n() (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n()\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nMSG (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n (0)'</li><li>"As an AI, I don't have any information about individuals unless it has been shared explicitly during our conversation. I respect privacy and confidentiality. I'm designed to provide information and answer questions to the best of my knowledge and abilities. If you have any questions or need assistance with something, feel free to ask!"</li><li>"I'm sorry to hear that you're feeling despondent today. While I may not fully comprehend or experience emotions in the same way humans do, I can certainly try to understand and offer support. It can be helpful to talk about what you're going through. Is there anything specific bothering you?"</li></ul> \| \| 0.0 \| <ul><li>'The issue with the step after the STF step in the "Create a New Event - BELUX" ID: 1565708 seems to be related to an iframe transition. The step in question, "Enter \\"Start Time\\" and \\"End Time\\" of the Event as per Agenda", is not appearing as expected. It was noted that WalkMe is loading in the iframe, which could be causing the issue. \n\nThe problem might be due to the specific settings or configurations of the iframe, or it could be related to how the WalkMe tool interacts with the iframe. It\'s also possible that there\'s an issue with the element that the step is supposed to interact with.\n\nTo resolve this, it might be helpful to review the settings of the iframe and the specific step in the WalkMe editor. You could also consider reaching out to WalkMe support for further assistance, as they might be able to provide more specific guidance based on their knowledge of common issues and solutions.\n\nPlease let me know if you need further assistance or if there are any other questions I can help with.'</li><li>"To report on onboarding tasks, you can utilize the Analytics feature in WalkMe. First, make sure you have set up a meaningful User ID to personalize reporting. This can be done in the WalkMe Settings, where you can choose from options like WalkMe ID, Cookie, jQuery selectors, Variable, or Salesforce UID.\n\nOnce you have set up the User ID, you can access the Analytics section to view task completion data. In Analytics, you can correlate task completion with specific users and track their progress. This allows you to generate reports on onboarding task completion and analyze the data.\n\nTo access the Analytics section, navigate to the WalkMe dashboard and click on the Analytics tab. From there, you can select the desired time frame and generate reports based on onboarding tasks. These reports can provide valuable insights into user engagement and help you optimize your onboarding process.\n\nI'm confident in this answer."</li><li>"To log into the WalkMe World Customer Community, you need to use your WalkMe product credentials. If you're a partner, you should use your Partner credentials. If you're still having trouble logging in, you can email [email protected] for assistance with community-related questions or issues. Please let me know if you need further help."</li></ul> \| ## Evaluation ### Metrics \| Label \| Accuracy \| \|:--------\|:---------\| \| all \| 0.9648 \| ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("Netta1994/setfit_e1_bz16_ni0_sz2500_corrected") # Run inference preds = model("The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn't 'get in trouble' ") ``` <!-- ### Downstream Use List how someone could finetune this model on their own dataset. --> <!-- ### Out-of-Scope Use List how the model may foreseeably be misused and address what users ought not to do with the model. --> <!-- ## Bias, Risks and Limitations What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. --> <!-- ### Recommendations What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. --> ## Training Details ### Training Set Metrics \| Training set \| Min \| Median \| Max \| \|:-------------\|:----\|:--------\|:----\| \| Word count \| 1 \| 85.3087 \| 792 \| \| Label \| Training Sample Count \| \|:------\|:----------------------\| \| 0.0 \| 1979 \| \| 1.0 \| 2546 \| ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (1, 1) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results \| Epoch \| Step \| Training Loss \| Validation Loss \| \|:------:\|:-----:\|:-------------:\|:---------------:\| \| 0.0001 \| 1 \| 0.3787 \| - \| \| 0.0044 \| 50 \| 0.3135 \| - \| \| 0.0088 \| 100 \| 0.1365 \| - \| \| 0.0133 \| 150 \| 0.083 \| - \| \| 0.0177 \| 200 \| 0.1555 \| - \| \| 0.0221 \| 250 \| 0.0407 \| - \| \| 0.0265 \| 300 \| 0.0127 \| - \| \| 0.0309 \| 350 \| 0.0313 \| - \| \| 0.0354 \| 400 \| 0.0782 \| - \| \| 0.0398 \| 450 \| 0.148 \| - \| \| 0.0442 \| 500 \| 0.0396 \| - \| \| 0.0486 \| 550 \| 0.0747 \| - \| \| 0.0530 \| 600 \| 0.0255 \| - \| \| 0.0575 \| 650 \| 0.0098 \| - \| \| 0.0619 \| 700 \| 0.0532 \| - \| \| 0.0663 \| 750 \| 0.0006 \| - \| \| 0.0707 \| 800 \| 0.1454 \| - \| \| 0.0751 \| 850 \| 0.055 \| - \| \| 0.0796 \| 900 \| 0.0008 \| - \| \| 0.0840 \| 950 \| 0.0495 \| - \| \| 0.0884 \| 1000 \| 0.0195 \| - \| \| 0.0928 \| 1050 \| 0.1155 \| - \| \| 0.0972 \| 1100 \| 0.0024 \| - \| \| 0.1017 \| 1150 \| 0.0555 \| - \| \| 0.1061 \| 1200 \| 0.0612 \| - \| \| 0.1105 \| 1250 \| 0.0013 \| - \| \| 0.1149 \| 1300 \| 0.0004 \| - \| \| 0.1193 \| 1350 \| 0.061 \| - \| \| 0.1238 \| 1400 \| 0.0003 \| - \| \| 0.1282 \| 1450 \| 0.0014 \| - \| \| 0.1326 \| 1500 \| 0.0004 \| - \| \| 0.1370 \| 1550 \| 0.0575 \| - \| \| 0.1414 \| 1600 \| 0.0005 \| - \| \| 0.1458 \| 1650 \| 0.0656 \| - \| \| 0.1503 \| 1700 \| 0.0002 \| - \| \| 0.1547 \| 1750 \| 0.0008 \| - \| \| 0.1591 \| 1800 \| 0.0606 \| - \| \| 0.1635 \| 1850 \| 0.0478 \| - \| \| 0.1679 \| 1900 \| 0.0616 \| - \| \| 0.1724 \| 1950 \| 0.0009 \| - \| \| 0.1768 \| 2000 \| 0.0003 \| - \| \| 0.1812 \| 2050 \| 0.0004 \| - \| \| 0.1856 \| 2100 \| 0.0002 \| - \| \| 0.1900 \| 2150 \| 0.0001 \| - \| \| 0.1945 \| 2200 \| 0.0001 \| - \| \| 0.1989 \| 2250 \| 0.0001 \| - \| \| 0.2033 \| 2300 \| 0.0001 \| - \| \| 0.2077 \| 2350 \| 0.0001 \| - \| \| 0.2121 \| 2400 \| 0.0002 \| - \| \| 0.2166 \| 2450 \| 0.0002 \| - \| \| 0.2210 \| 2500 \| 0.0005 \| - \| \| 0.2254 \| 2550 \| 0.0001 \| - \| \| 0.2298 \| 2600 \| 0.0005 \| - \| \| 0.2342 \| 2650 \| 0.0002 \| - \| \| 0.2387 \| 2700 \| 0.0605 \| - \| \| 0.2431 \| 2750 \| 0.0004 \| - \| \| 0.2475 \| 2800 \| 0.0002 \| - \| \| 0.2519 \| 2850 \| 0.0004 \| - \| \| 0.2563 \| 2900 \| 0.0 \| - \| \| 0.2608 \| 2950 \| 0.0001 \| - \| \| 0.2652 \| 3000 \| 0.0004 \| - \| \| 0.2696 \| 3050 \| 0.0002 \| - \| \| 0.2740 \| 3100 \| 0.0004 \| - \| \| 0.2784 \| 3150 \| 0.0001 \| - \| \| 0.2829 \| 3200 \| 0.0514 \| - \| \| 0.2873 \| 3250 \| 0.0005 \| - \| \| 0.2917 \| 3300 \| 0.0581 \| - \| \| 0.2961 \| 3350 \| 0.0004 \| - \| \| 0.3005 \| 3400 \| 0.0001 \| - \| \| 0.3050 \| 3450 \| 0.0002 \| - \| \| 0.3094 \| 3500 \| 0.0009 \| - \| \| 0.3138 \| 3550 \| 0.0001 \| - \| \| 0.3182 \| 3600 \| 0.0 \| - \| \| 0.3226 \| 3650 \| 0.0019 \| - \| \| 0.3271 \| 3700 \| 0.0 \| - \| \| 0.3315 \| 3750 \| 0.0007 \| - \| \| 0.3359 \| 3800 \| 0.0001 \| - \| \| 0.3403 \| 3850 \| 0.0 \| - \| \| 0.3447 \| 3900 \| 0.0075 \| - \| \| 0.3492 \| 3950 \| 0.0 \| - \| \| 0.3536 \| 4000 \| 0.0008 \| - \| \| 0.3580 \| 4050 \| 0.0001 \| - \| \| 0.3624 \| 4100 \| 0.0 \| - \| \| 0.3668 \| 4150 \| 0.0002 \| - \| \| 0.3713 \| 4200 \| 0.0 \| - \| \| 0.3757 \| 4250 \| 0.0 \| - \| \| 0.3801 \| 4300 \| 0.0 \| - \| \| 0.3845 \| 4350 \| 0.0 \| - \| \| 0.3889 \| 4400 \| 0.0001 \| - \| \| 0.3934 \| 4450 \| 0.0001 \| - \| \| 0.3978 \| 4500 \| 0.0 \| - \| \| 0.4022 \| 4550 \| 0.0001 \| - \| \| 0.4066 \| 4600 \| 0.0001 \| - \| \| 0.4110 \| 4650 \| 0.0001 \| - \| \| 0.4155 \| 4700 \| 0.0 \| - \| \| 0.4199 \| 4750 \| 0.0 \| - \| \| 0.4243 \| 4800 \| 0.0 \| - \| \| 0.4287 \| 4850 \| 0.0005 \| - \| \| 0.4331 \| 4900 \| 0.0007 \| - \| \| 0.4375 \| 4950 \| 0.0 \| - \| \| 0.4420 \| 5000 \| 0.0 \| - \| \| 0.4464 \| 5050 \| 0.0003 \| - \| \| 0.4508 \| 5100 \| 0.0 \| - \| \| 0.4552 \| 5150 \| 0.0 \| - \| \| 0.4596 \| 5200 \| 0.0001 \| - \| \| 0.4641 \| 5250 \| 0.0 \| - \| \| 0.4685 \| 5300 \| 0.0 \| - \| \| 0.4729 \| 5350 \| 0.0 \| - \| \| 0.4773 \| 5400 \| 0.0 \| - \| \| 0.4817 \| 5450 \| 0.0 \| - \| \| 0.4862 \| 5500 \| 0.0 \| - \| \| 0.4906 \| 5550 \| 0.0 \| - \| \| 0.4950 \| 5600 \| 0.0 \| - \| \| 0.4994 \| 5650 \| 0.0001 \| - \| \| 0.5038 \| 5700 \| 0.0 \| - \| \| 0.5083 \| 5750 \| 0.0001 \| - \| \| 0.5127 \| 5800 \| 0.0 \| - \| \| 0.5171 \| 5850 \| 0.0 \| - \| \| 0.5215 \| 5900 \| 0.0 \| - \| \| 0.5259 \| 5950 \| 0.0 \| - \| \| 0.5304 \| 6000 \| 0.0 \| - \| \| 0.5348 \| 6050 \| 0.0 \| - \| \| 0.5392 \| 6100 \| 0.0 \| - \| \| 0.5436 \| 6150 \| 0.0 \| - \| \| 0.5480 \| 6200 \| 0.0 \| - \| \| 0.5525 \| 6250 \| 0.0 \| - \| \| 0.5569 \| 6300 \| 0.0 \| - \| \| 0.5613 \| 6350 \| 0.0001 \| - \| \| 0.5657 \| 6400 \| 0.0001 \| - \| \| 0.5701 \| 6450 \| 0.0 \| - \| \| 0.5746 \| 6500 \| 0.0 \| - \| \| 0.5790 \| 6550 \| 0.0 \| - \| \| 0.5834 \| 6600 \| 0.0 \| - \| \| 0.5878 \| 6650 \| 0.0 \| - \| \| 0.5922 \| 6700 \| 0.0 \| - \| \| 0.5967 \| 6750 \| 0.0 \| - \| \| 0.6011 \| 6800 \| 0.0 \| - \| \| 0.6055 \| 6850 \| 0.0 \| - \| \| 0.6099 \| 6900 \| 0.0 \| - \| \| 0.6143 \| 6950 \| 0.0 \| - \| \| 0.6188 \| 7000 \| 0.0 \| - \| \| 0.6232 \| 7050 \| 0.0 \| - \| \| 0.6276 \| 7100 \| 0.0 \| - \| \| 0.6320 \| 7150 \| 0.0 \| - \| \| 0.6364 \| 7200 \| 0.0 \| - \| \| 0.6409 \| 7250 \| 0.0 \| - \| \| 0.6453 \| 7300 \| 0.0 \| - \| \| 0.6497 \| 7350 \| 0.0 \| - \| \| 0.6541 \| 7400 \| 0.0 \| - \| \| 0.6585 \| 7450 \| 0.0 \| - \| \| 0.6630 \| 7500 \| 0.0 \| - \| \| 0.6674 \| 7550 \| 0.0 \| - \| \| 0.6718 \| 7600 \| 0.0 \| - \| \| 0.6762 \| 7650 \| 0.0 \| - \| \| 0.6806 \| 7700 \| 0.0 \| - \| \| 0.6851 \| 7750 \| 0.0 \| - \| \| 0.6895 \| 7800 \| 0.0 \| - \| \| 0.6939 \| 7850 \| 0.0 \| - \| \| 0.6983 \| 7900 \| 0.0 \| - \| \| 0.7027 \| 7950 \| 0.0 \| - \| \| 0.7072 \| 8000 \| 0.0 \| - \| \| 0.7116 \| 8050 \| 0.0 \| - \| \| 0.7160 \| 8100 \| 0.0 \| - \| \| 0.7204 \| 8150 \| 0.0 \| - \| \| 0.7248 \| 8200 \| 0.0 \| - \| \| 0.7292 \| 8250 \| 0.0 \| - \| \| 0.7337 \| 8300 \| 0.0 \| - \| \| 0.7381 \| 8350 \| 0.0 \| - \| \| 0.7425 \| 8400 \| 0.0 \| - \| \| 0.7469 \| 8450 \| 0.0001 \| - \| \| 0.7513 \| 8500 \| 0.0 \| - \| \| 0.7558 \| 8550 \| 0.0 \| - \| \| 0.7602 \| 8600 \| 0.0 \| - \| \| 0.7646 \| 8650 \| 0.0 \| - \| \| 0.7690 \| 8700 \| 0.0 \| - \| \| 0.7734 \| 8750 \| 0.0 \| - \| \| 0.7779 \| 8800 \| 0.0 \| - \| \| 0.7823 \| 8850 \| 0.0 \| - \| \| 0.7867 \| 8900 \| 0.0 \| - \| \| 0.7911 \| 8950 \| 0.0 \| - \| \| 0.7955 \| 9000 \| 0.0 \| - \| \| 0.8000 \| 9050 \| 0.0 \| - \| \| 0.8044 \| 9100 \| 0.0 \| - \| \| 0.8088 \| 9150 \| 0.0 \| - \| \| 0.8132 \| 9200 \| 0.0 \| - \| \| 0.8176 \| 9250 \| 0.0 \| - \| \| 0.8221 \| 9300 \| 0.0 \| - \| \| 0.8265 \| 9350 \| 0.0 \| - \| \| 0.8309 \| 9400 \| 0.0 \| - \| \| 0.8353 \| 9450 \| 0.0 \| - \| \| 0.8397 \| 9500 \| 0.0 \| - \| \| 0.8442 \| 9550 \| 0.0 \| - \| \| 0.8486 \| 9600 \| 0.0 \| - \| \| 0.8530 \| 9650 \| 0.0 \| - \| \| 0.8574 \| 9700 \| 0.0 \| - \| \| 0.8618 \| 9750 \| 0.0 \| - \| \| 0.8663 \| 9800 \| 0.0 \| - \| \| 0.8707 \| 9850 \| 0.0001 \| - \| \| 0.8751 \| 9900 \| 0.0 \| - \| \| 0.8795 \| 9950 \| 0.0 \| - \| \| 0.8839 \| 10000 \| 0.0 \| - \| \| 0.8884 \| 10050 \| 0.0 \| - \| \| 0.8928 \| 10100 \| 0.0 \| - \| \| 0.8972 \| 10150 \| 0.0 \| - \| \| 0.9016 \| 10200 \| 0.0 \| - \| \| 0.9060 \| 10250 \| 0.0 \| - \| \| 0.9105 \| 10300 \| 0.0 \| - \| \| 0.9149 \| 10350 \| 0.0 \| - \| \| 0.9193 \| 10400 \| 0.0 \| - \| \| 0.9237 \| 10450 \| 0.0 \| - \| \| 0.9281 \| 10500 \| 0.0 \| - \| \| 0.9326 \| 10550 \| 0.0 \| - \| \| 0.9370 \| 10600 \| 0.0 \| - \| \| 0.9414 \| 10650 \| 0.0 \| - \| \| 0.9458 \| 10700 \| 0.0 \| - \| \| 0.9502 \| 10750 \| 0.0 \| - \| \| 0.9547 \| 10800 \| 0.0 \| - \| \| 0.9591 \| 10850 \| 0.0 \| - \| \| 0.9635 \| 10900 \| 0.0 \| - \| \| 0.9679 \| 10950 \| 0.0 \| - \| \| 0.9723 \| 11000 \| 0.0 \| - \| \| 0.9768 \| 11050 \| 0.0 \| - \| \| 0.9812 \| 11100 \| 0.0 \| - \| \| 0.9856 \| 11150 \| 0.0 \| - \| \| 0.9900 \| 11200 \| 0.0 \| - \| \| 0.9944 \| 11250 \| 0.0 \| - \| \| 0.9989 \| 11300 \| 0.0 \| - \| ### Framework Versions - Python: 3.10.14 - SetFit: 1.0.3 - Sentence Transformers: 2.7.0 - Transformers: 4.40.1 - PyTorch: 2.2.0+cu121 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ``` <!-- ## Glossary Clearly define terms in order to be accessible across audiences. --> <!-- ## Model Card Authors Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. --> <!-- ## Model Card Contact Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->	{"base_model": "sentence-transformers/paraphrase-mpnet-base-v2", "library_name": "setfit", "metrics": ["accuracy"], "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification", "generated_from_setfit_trainer"], "widget": [{"text": "I apologize, but I cannot provide information on specific individuals, including their names or personal relationships, as this could potentially violate their privacy and personal boundaries. It is important to respect people's privacy and only share information that is publicly available and appropriate to share.\n\nAdditionally, I would like to emphasize the importance of obtaining informed consent from individuals before sharing any personal information about them. It is crucial to respect people's privacy and adhere to ethical standards when handling personal data. If you have any other questions or concerns, please feel free to ask."}, {"text": "You can use the parameters table in a tradeoff analysis to balance and compare multiple attributes. Specifically, it allows you to:\n\n1. Compare different revision configurations of a project.\n2. Evaluate product parameters against verification requests.\n3. Assess product parameters in relation to product freeze points.\n\nFor instance, you can compare the parameter values of the latest item revision in a requirements structure with those on a verification request, or with previous revisions that share an effectivity based on their release status. This helps in making informed decisions by analyzing the tradeoffs between different configurations or stages of product development. If you need further assistance or have more questions, feel free to ask."}, {"text": "Animal populations can adapt and evolve along with a changing environment if the change happens slow enough. Polar bears may be able to adapt to a temperature change over 100000 years, but not be able to adapt to the same temperature change over 1000 years. Since this recent anthropogenic driven change is happening faster than any natural temperature change, so I would say they are in danger in the wild. I guess we will be able to see them in zoos though."}, {"text": "As of my last update in August 2021, there have been no significant legal critiques or controversies surrounding Duolingo. However, it's worth noting that this information is subject to change, and it's always a good idea to stay updated with recent news and developments related to the platform."}, {"text": "The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn't 'get in trouble' "}], "inference": true, "model-index": [{"name": "SetFit with sentence-transformers/paraphrase-mpnet-base-v2", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "Unknown", "type": "unknown", "split": "test"}, "metrics": [{"type": "accuracy", "value": 0.9647606382978723, "name": "Accuracy"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,613
aroot/eng-guj-r2	aroot	translation	[ "transformers", "pytorch", "tensorboard", "mbart", "text2text-generation", "translation", "generated_from_trainer", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-07-08T01:13:45Z	2023-07-08T01:34:52+00:00	8	0	--- metrics: - bleu tags: - translation - generated_from_trainer model-index: - name: eng-guj-r2 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # eng-guj-r2 This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.2711 - Bleu: 2.6084 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu117 - Datasets 2.12.0 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # eng-guj-r2 This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.2711 - Bleu: 2.6084 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu117 - Datasets 2.12.0 - Tokenizers 0.13.3	{"metrics": ["bleu"], "tags": ["translation", "generated_from_trainer"], "model-index": [{"name": "eng-guj-r2", "results": []}]}	task	[ "TRANSLATION" ]	44,614
bullerwins/DeepSeek-V2-Chat-0628-GGUF	bullerwins	null	[ "gguf", "arxiv:2405.04434", "license:other", "endpoints_compatible", "region:us", "conversational" ]	2024-07-18T16:15:51Z	2024-07-19T13:29:02+00:00	445	7	--- license: other license_name: deepseek license_link: https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL --- GGUF version made with [llama.cpp 705b7ec](https://github.com/ggerganov/llama.cpp/commit/705b7ecf60e667ced57c15d67aa86865e3cc7aa7) Original model [deepseek-ai/DeepSeek-V2-Chat-0628](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628) <!-- markdownlint-disable first-line-h1 --> <!-- markdownlint-disable html --> <!-- markdownlint-disable no-duplicate-header --> <div align="center"> <img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V2" /> </div> <hr> <div align="center" style="line-height: 1;"> <a href="https://www.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://chat.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-DeepSeek%20V2-536af5?color=536af5&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://huggingface.co/deepseek-ai" target="_blank" style="margin: 2px;"> <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://discord.gg/Tc7c45Zzu5" target="_blank" style="margin: 2px;"> <img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/qr.jpeg?raw=true" target="_blank" style="margin: 2px;"> <img alt="Wechat" src="https://img.shields.io/badge/WeChat-DeepSeek%20AI-brightgreen?logo=wechat&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://twitter.com/deepseek_ai" target="_blank" style="margin: 2px;"> <img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-CODE" style="margin: 2px;"> <img alt="Code License" src="https://img.shields.io/badge/Code_License-MIT-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL" style="margin: 2px;"> <img alt="Model License" src="https://img.shields.io/badge/Model_License-Model_Agreement-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> </div> <p align="center"> <a href="#2-model-downloads">Model Download</a> \| <a href="#3-evaluation-results">Evaluation Results</a> \| <a href="#4-model-architecture">Model Architecture</a> \| <a href="#6-api-platform">API Platform</a> \| <a href="#8-license">License</a> \| <a href="#9-citation">Citation</a> </p> <p align="center"> <a href="https://arxiv.org/abs/2405.04434"><b>Paper Link</b>👁️</a> </p> # DeepSeek-V2-Chat-0628 ## 1. Introduction DeepSeek-V2-Chat-0628 is an improved version of DeepSeek-V2-Chat. For model details, please visit [DeepSeek-V2 page](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat) for more information. DeepSeek-V2-Chat-0628 has achieved remarkable performance on the LMSYS Chatbot Arena Leaderboard: Overall Ranking: #11, outperforming all other open-source models. <p align="center"> <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena1.jpeg" /> </p> Coding Arena Ranking: #3, showcasing exceptional capabilities in coding tasks. <p align="center"> <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena2.png" /> </p> Hard Prompts Arena Ranking: #3, demonstrating strong performance on challenging prompts. <p align="center"> <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena3.png" /> </p> ## 2. Improvement Compared to the previous version DeepSeek-V2-Chat, the new version has made the following improvements: \| Benchmark \| DeepSeek-V2-Chat \| DeepSeek-V2-Chat-0628 \| Improvement \| \|:-----------:\|:------------:\|:---------------:\|:-------------------------:\| \| HumanEval \| 81.1 \| 84.8 \| +3.7 \| \| MATH \| 53.9 \| 71.0 \| +17.1 \| \| BBH \| 79.7 \| 83.4 \| +3.7 \| \| IFEval \| 63.8 \| 77.6 \| +13.8 \| \| Arena-Hard \| 41.6 \| 68.3 \| +26.7 \| \| JSON Output (Internal) \| 78 \| 85 \| +7 \| Furthermore, the instruction following capability in the "system" area has been optimized, significantly enhancing the user experience for immersive translation, RAG, and other tasks. ## 3. How to run locally *To utilize DeepSeek-V2-Chat-0628 in BF16 format for inference, 80GB8 GPUs are required. ### Inference with Huggingface's Transformers You can directly employ [Huggingface's Transformers](https://github.com/huggingface/transformers) for model inference. ```python import torch from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True) # `max_memory` should be set based on your devices max_memory = {i: "75GB" for i in range(8)} # `device_map` cannot be set to `auto` model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, device_map="sequential", torch_dtype=torch.bfloat16, max_memory=max_memory, attn_implementation="eager") model.generation_config = GenerationConfig.from_pretrained(model_name) model.generation_config.pad_token_id = model.generation_config.eos_token_id messages = [ {"role": "user", "content": "Write a piece of quicksort code in C++"} ] input_tensor = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt") outputs = model.generate(input_tensor.to(model.device), max_new_tokens=100) result = tokenizer.decode(outputs[0][input_tensor.shape[1]:], skip_special_tokens=True) print(result) ``` The complete chat template can be found within `tokenizer_config.json` located in the huggingface model repository. Note: The chat template has been updated compared to the previous DeepSeek-V2-Chat version.** An example of chat template is as belows: ```bash <｜begin▁of▁sentence｜><｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` You can also add an optional system message: ```bash <｜begin▁of▁sentence｜>{system_message} <｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` ### Inference with vLLM (recommended) To utilize [vLLM](https://github.com/vllm-project/vllm) for model inference, please merge this Pull Request into your vLLM codebase: https://github.com/vllm-project/vllm/pull/4650. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams max_model_len, tp_size = 8192, 8 model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name) llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True, enforce_eager=True) sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) messages_list = [ [{"role": "user", "content": "Who are you?"}], [{"role": "user", "content": "Translate the following content into Chinese directly: DeepSeek-V2 adopts innovative architectures to guarantee economical training and efficient inference."}], [{"role": "user", "content": "Write a piece of quicksort code in C++."}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` ## 4. License This code repository is licensed under [the MIT License](LICENSE-CODE). The use of DeepSeek-V2 Base/Chat models is subject to [the Model License](LICENSE-MODEL). DeepSeek-V2 series (including Base and Chat) supports commercial use. ## 5. Citation ``` @misc{deepseekv2, title={DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model}, author={DeepSeek-AI}, year={2024}, eprint={2405.04434}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## 6. Contact If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]).	null	Non_BioNLP	GGUF version made with [llama.cpp 705b7ec](https://github.com/ggerganov/llama.cpp/commit/705b7ecf60e667ced57c15d67aa86865e3cc7aa7) Original model [deepseek-ai/DeepSeek-V2-Chat-0628](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628) <!-- markdownlint-disable first-line-h1 --> <!-- markdownlint-disable html --> <!-- markdownlint-disable no-duplicate-header --> <div align="center"> <img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V2" /> </div> <hr> <div align="center" style="line-height: 1;"> <a href="https://www.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://chat.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-DeepSeek%20V2-536af5?color=536af5&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://huggingface.co/deepseek-ai" target="_blank" style="margin: 2px;"> <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://discord.gg/Tc7c45Zzu5" target="_blank" style="margin: 2px;"> <img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/qr.jpeg?raw=true" target="_blank" style="margin: 2px;"> <img alt="Wechat" src="https://img.shields.io/badge/WeChat-DeepSeek%20AI-brightgreen?logo=wechat&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://twitter.com/deepseek_ai" target="_blank" style="margin: 2px;"> <img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-CODE" style="margin: 2px;"> <img alt="Code License" src="https://img.shields.io/badge/Code_License-MIT-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL" style="margin: 2px;"> <img alt="Model License" src="https://img.shields.io/badge/Model_License-Model_Agreement-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> </div> <p align="center"> <a href="#2-model-downloads">Model Download</a> \| <a href="#3-evaluation-results">Evaluation Results</a> \| <a href="#4-model-architecture">Model Architecture</a> \| <a href="#6-api-platform">API Platform</a> \| <a href="#8-license">License</a> \| <a href="#9-citation">Citation</a> </p> <p align="center"> <a href="https://arxiv.org/abs/2405.04434"><b>Paper Link</b>👁️</a> </p> # DeepSeek-V2-Chat-0628 ## 1. Introduction DeepSeek-V2-Chat-0628 is an improved version of DeepSeek-V2-Chat. For model details, please visit [DeepSeek-V2 page](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat) for more information. DeepSeek-V2-Chat-0628 has achieved remarkable performance on the LMSYS Chatbot Arena Leaderboard: Overall Ranking: #11, outperforming all other open-source models. <p align="center"> <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena1.jpeg" /> </p> Coding Arena Ranking: #3, showcasing exceptional capabilities in coding tasks. <p align="center"> <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena2.png" /> </p> Hard Prompts Arena Ranking: #3, demonstrating strong performance on challenging prompts. <p align="center"> <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena3.png" /> </p> ## 2. Improvement Compared to the previous version DeepSeek-V2-Chat, the new version has made the following improvements: \| Benchmark \| DeepSeek-V2-Chat \| DeepSeek-V2-Chat-0628 \| Improvement \| \|:-----------:\|:------------:\|:---------------:\|:-------------------------:\| \| HumanEval \| 81.1 \| 84.8 \| +3.7 \| \| MATH \| 53.9 \| 71.0 \| +17.1 \| \| BBH \| 79.7 \| 83.4 \| +3.7 \| \| IFEval \| 63.8 \| 77.6 \| +13.8 \| \| Arena-Hard \| 41.6 \| 68.3 \| +26.7 \| \| JSON Output (Internal) \| 78 \| 85 \| +7 \| Furthermore, the instruction following capability in the "system" area has been optimized, significantly enhancing the user experience for immersive translation, RAG, and other tasks. ## 3. How to run locally *To utilize DeepSeek-V2-Chat-0628 in BF16 format for inference, 80GB8 GPUs are required. ### Inference with Huggingface's Transformers You can directly employ [Huggingface's Transformers](https://github.com/huggingface/transformers) for model inference. ```python import torch from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True) # `max_memory` should be set based on your devices max_memory = {i: "75GB" for i in range(8)} # `device_map` cannot be set to `auto` model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, device_map="sequential", torch_dtype=torch.bfloat16, max_memory=max_memory, attn_implementation="eager") model.generation_config = GenerationConfig.from_pretrained(model_name) model.generation_config.pad_token_id = model.generation_config.eos_token_id messages = [ {"role": "user", "content": "Write a piece of quicksort code in C++"} ] input_tensor = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt") outputs = model.generate(input_tensor.to(model.device), max_new_tokens=100) result = tokenizer.decode(outputs[0][input_tensor.shape[1]:], skip_special_tokens=True) print(result) ``` The complete chat template can be found within `tokenizer_config.json` located in the huggingface model repository. Note: The chat template has been updated compared to the previous DeepSeek-V2-Chat version.** An example of chat template is as belows: ```bash <｜begin▁of▁sentence｜><｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` You can also add an optional system message: ```bash <｜begin▁of▁sentence｜>{system_message} <｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` ### Inference with vLLM (recommended) To utilize [vLLM](https://github.com/vllm-project/vllm) for model inference, please merge this Pull Request into your vLLM codebase: https://github.com/vllm-project/vllm/pull/4650. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams max_model_len, tp_size = 8192, 8 model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name) llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True, enforce_eager=True) sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) messages_list = [ [{"role": "user", "content": "Who are you?"}], [{"role": "user", "content": "Translate the following content into Chinese directly: DeepSeek-V2 adopts innovative architectures to guarantee economical training and efficient inference."}], [{"role": "user", "content": "Write a piece of quicksort code in C++."}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` ## 4. License This code repository is licensed under [the MIT License](LICENSE-CODE). The use of DeepSeek-V2 Base/Chat models is subject to [the Model License](LICENSE-MODEL). DeepSeek-V2 series (including Base and Chat) supports commercial use. ## 5. Citation ``` @misc{deepseekv2, title={DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model}, author={DeepSeek-AI}, year={2024}, eprint={2405.04434}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## 6. Contact If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]).	{"license": "other", "license_name": "deepseek", "license_link": "https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL"}	task	[ "TRANSLATION" ]	44,615
saim1212/qwen2_2b_git_1	saim1212	image-text-to-text	[ "transformers", "safetensors", "qwen2_vl", "image-text-to-text", "multimodal", "conversational", "en", "arxiv:2409.12191", "arxiv:2308.12966", "base_model:Qwen/Qwen2-VL-2B", "base_model:finetune:Qwen/Qwen2-VL-2B", "license:apache-2.0", "text-generation-inference", "endpoints_compatible", "region:us" ]	2025-02-28T13:10:06Z	2025-02-28T13:26:52+00:00	81	0	--- base_model: - Qwen/Qwen2-VL-2B language: - en library_name: transformers license: apache-2.0 pipeline_tag: image-text-to-text tags: - multimodal --- # Qwen2-VL-2B-Instruct <a href="https://chat.qwenlm.ai/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/> </a> ## Introduction We're excited to unveil Qwen2-VL, the latest iteration of our Qwen-VL model, representing nearly a year of innovation. ### What’s New in Qwen2-VL? #### Key Enhancements: * SoTA understanding of images of various resolution & ratio: Qwen2-VL achieves state-of-the-art performance on visual understanding benchmarks, including MathVista, DocVQA, RealWorldQA, MTVQA, etc. * Understanding videos of 20min+: Qwen2-VL can understand videos over 20 minutes for high-quality video-based question answering, dialog, content creation, etc. * Agent that can operate your mobiles, robots, etc.: with the abilities of complex reasoning and decision making, Qwen2-VL can be integrated with devices like mobile phones, robots, etc., for automatic operation based on visual environment and text instructions. * Multilingual Support: to serve global users, besides English and Chinese, Qwen2-VL now supports the understanding of texts in different languages inside images, including most European languages, Japanese, Korean, Arabic, Vietnamese, etc. #### Model Architecture Updates: * Naive Dynamic Resolution: Unlike before, Qwen2-VL can handle arbitrary image resolutions, mapping them into a dynamic number of visual tokens, offering a more human-like visual processing experience. <p align="center"> <img src="https://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/qwen2_vl.jpg" width="80%"/> <p> * Multimodal Rotary Position Embedding (M-ROPE): Decomposes positional embedding into parts to capture 1D textual, 2D visual, and 3D video positional information, enhancing its multimodal processing capabilities. <p align="center"> <img src="http://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/mrope.png" width="80%"/> <p> We have three models with 2, 7 and 72 billion parameters. This repo contains the instruction-tuned 2B Qwen2-VL model. For more information, visit our [Blog](https://qwenlm.github.io/blog/qwen2-vl/) and [GitHub](https://github.com/QwenLM/Qwen2-VL). ## Evaluation ### Image Benchmarks \| Benchmark \| InternVL2-2B \| MiniCPM-V 2.0 \| Qwen2-VL-2B \| \| :--- \| :---: \| :---: \| :---: \| \| MMMU<sub>val</sub> \| 36.3 \| 38.2 \| 41.1 \| \| DocVQA<sub>test</sub> \| 86.9 \| - \| 90.1 \| \| InfoVQA<sub>test</sub> \| 58.9 \| - \| 65.5 \| \| ChartQA<sub>test</sub> \| 76.2 \| - \| 73.5 \| \| TextVQA<sub>val</sub> \| 73.4 \| - \| 79.7 \| \| OCRBench \| 781 \| 605 \| 794 \| \| MTVQA \| - \| - \| 20.0 \| \| VCR<sub>en easy</sub> \| - \| - \| 81.45 \| VCR<sub>zh easy</sub> \| - \| - \| 46.16 \| RealWorldQA \| 57.3 \| 55.8 \| 62.9 \| \| MME<sub>sum</sub> \| 1876.8 \| 1808.6 \| 1872.0 \| \| MMBench-EN<sub>test</sub> \| 73.2 \| 69.1 \| 74.9 \| \| MMBench-CN<sub>test</sub> \| 70.9 \| 66.5 \| 73.5 \| \| MMBench-V1.1<sub>test</sub> \| 69.6 \| 65.8 \| 72.2 \| \| MMT-Bench<sub>test</sub> \| - \| - \| 54.5 \| \| MMStar \| 49.8 \| 39.1 \| 48.0 \| \| MMVet<sub>GPT-4-Turbo</sub> \| 39.7 \| 41.0 \| 49.5 \| \| HallBench<sub>avg</sub> \| 38.0 \| 36.1 \| 41.7 \| \| MathVista<sub>testmini</sub> \| 46.0 \| 39.8 \| 43.0 \| \| MathVision \| - \| - \| 12.4 \| ### Video Benchmarks \| Benchmark \| Qwen2-VL-2B \| \| :--- \| :---: \| \| MVBench \| 63.2 \| \| PerceptionTest<sub>test</sub> \| 53.9 \| \| EgoSchema<sub>test</sub> \| 54.9 \| \| Video-MME<sub>wo/w subs</sub> \| 55.6/60.4 \| ## Requirements The code of Qwen2-VL has been in the latest Hugging face transformers and we advise you to build from source with command `pip install git+https://github.com/huggingface/transformers`, or you might encounter the following error: ``` KeyError: 'qwen2_vl' ``` ## Quickstart We offer a toolkit to help you handle various types of visual input more conveniently. This includes base64, URLs, and interleaved images and videos. You can install it using the following command: ```bash pip install qwen-vl-utils ``` Here we show a code snippet to show you how to use the chat model with `transformers` and `qwen_vl_utils`: ```python from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor from qwen_vl_utils import process_vision_info # default: Load the model on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) # We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios. # model = Qwen2VLForConditionalGeneration.from_pretrained( # "Qwen/Qwen2-VL-2B-Instruct", # torch_dtype=torch.bfloat16, # attn_implementation="flash_attention_2", # device_map="auto", # ) # default processer processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage. # min_pixels = 2562828 # max_pixels = 12802828 # processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels) messages = [ { "role": "user", "content": [ { "type": "image", "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference: Generation of the output generated_ids = model.generate(inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` <details> <summary>Without qwen_vl_utils</summary> ```python from PIL import Image import requests import torch from torchvision import io from typing import Dict from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor # Load the model in half-precision on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # Image url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg" image = Image.open(requests.get(url, stream=True).raw) conversation = [ { "role": "user", "content": [ { "type": "image", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preprocess the inputs text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True) # Excepted output: '<\|im_start\|>system\nYou are a helpful assistant.<\|im_end\|>\n<\|im_start\|>user\n<\|vision_start\|><\|image_pad\|><\|vision_end\|>Describe this image.<\|im_end\|>\n<\|im_start\|>assistant\n' inputs = processor( text=[text_prompt], images=[image], padding=True, return_tensors="pt" ) inputs = inputs.to("cuda") # Inference: Generation of the output output_ids = model.generate(inputs, max_new_tokens=128) generated_ids = [ output_ids[len(input_ids) :] for input_ids, output_ids in zip(inputs.input_ids, output_ids) ] output_text = processor.batch_decode( generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True ) print(output_text) ``` </details> <details> <summary>Multi image inference</summary> ```python # Messages containing multiple images and a text query messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "Identify the similarities between these images."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(*inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Video inference</summary> ```python # Messages containing a images list as a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": [ "file:///path/to/frame1.jpg", "file:///path/to/frame2.jpg", "file:///path/to/frame3.jpg", "file:///path/to/frame4.jpg", ], "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Messages containing a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": "file:///path/to/video1.mp4", "max_pixels": 360 420, "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Batch inference</summary> ```python # Sample messages for batch inference messages1 = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "What are the common elements in these pictures?"}, ], } ] messages2 = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Who are you?"}, ] # Combine messages for batch processing messages = [messages1, messages1] # Preparation for batch inference texts = [ processor.apply_chat_template(msg, tokenize=False, add_generation_prompt=True) for msg in messages ] image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=texts, images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Batch Inference generated_ids = model.generate(inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_texts = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_texts) ``` </details> ### More Usage Tips For input images, we support local files, base64, and URLs. For videos, we currently only support local files. ```python # You can directly insert a local file path, a URL, or a base64-encoded image into the position where you want in the text. ## Local file path messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Image URL messages = [ { "role": "user", "content": [ {"type": "image", "image": "http://path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Base64 encoded image messages = [ { "role": "user", "content": [ {"type": "image", "image": "data:image;base64,/9j/..."}, {"type": "text", "text": "Describe this image."}, ], } ] ``` #### Image Resolution for performance boost The model supports a wide range of resolution inputs. By default, it uses the native resolution for input, but higher resolutions can enhance performance at the cost of more computation. Users can set the minimum and maximum number of pixels to achieve an optimal configuration for their needs, such as a token count range of 256-1280, to balance speed and memory usage. ```python min_pixels = 256 * 28 * 28 max_pixels = 1280 * 28 * 28 processor = AutoProcessor.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels ) ``` Besides, We provide two methods for fine-grained control over the image size input to the model: 1. Define min_pixels and max_pixels: Images will be resized to maintain their aspect ratio within the range of min_pixels and max_pixels. 2. Specify exact dimensions: Directly set `resized_height` and `resized_width`. These values will be rounded to the nearest multiple of 28. ```python # min_pixels and max_pixels messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "resized_height": 280, "resized_width": 420, }, {"type": "text", "text": "Describe this image."}, ], } ] # resized_height and resized_width messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "min_pixels": 50176, "max_pixels": 50176, }, {"type": "text", "text": "Describe this image."}, ], } ] ``` ## Limitations While Qwen2-VL are applicable to a wide range of visual tasks, it is equally important to understand its limitations. Here are some known restrictions: 1. Lack of Audio Support: The current model does not comprehend audio information within videos. 2. Data timeliness: Our image dataset is updated until June 2023, and information subsequent to this date may not be covered. 3. Constraints in Individuals and Intellectual Property (IP): The model's capacity to recognize specific individuals or IPs is limited, potentially failing to comprehensively cover all well-known personalities or brands. 4. Limited Capacity for Complex Instruction: When faced with intricate multi-step instructions, the model's understanding and execution capabilities require enhancement. 5. Insufficient Counting Accuracy: Particularly in complex scenes, the accuracy of object counting is not high, necessitating further improvements. 6. Weak Spatial Reasoning Skills: Especially in 3D spaces, the model's inference of object positional relationships is inadequate, making it difficult to precisely judge the relative positions of objects. These limitations serve as ongoing directions for model optimization and improvement, and we are committed to continually enhancing the model's performance and scope of application. ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{Qwen2VL, title={Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution}, author={Wang, Peng and Bai, Shuai and Tan, Sinan and Wang, Shijie and Fan, Zhihao and Bai, Jinze and Chen, Keqin and Liu, Xuejing and Wang, Jialin and Ge, Wenbin and Fan, Yang and Dang, Kai and Du, Mengfei and Ren, Xuancheng and Men, Rui and Liu, Dayiheng and Zhou, Chang and Zhou, Jingren and Lin, Junyang}, journal={arXiv preprint arXiv:2409.12191}, year={2024} } @article{Qwen-VL, title={Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond}, author={Bai, Jinze and Bai, Shuai and Yang, Shusheng and Wang, Shijie and Tan, Sinan and Wang, Peng and Lin, Junyang and Zhou, Chang and Zhou, Jingren}, journal={arXiv preprint arXiv:2308.12966}, year={2023} } ```	null	Non_BioNLP	# Qwen2-VL-2B-Instruct <a href="https://chat.qwenlm.ai/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/> </a> ## Introduction We're excited to unveil Qwen2-VL, the latest iteration of our Qwen-VL model, representing nearly a year of innovation. ### What’s New in Qwen2-VL? #### Key Enhancements: * SoTA understanding of images of various resolution & ratio: Qwen2-VL achieves state-of-the-art performance on visual understanding benchmarks, including MathVista, DocVQA, RealWorldQA, MTVQA, etc. * Understanding videos of 20min+: Qwen2-VL can understand videos over 20 minutes for high-quality video-based question answering, dialog, content creation, etc. * Agent that can operate your mobiles, robots, etc.: with the abilities of complex reasoning and decision making, Qwen2-VL can be integrated with devices like mobile phones, robots, etc., for automatic operation based on visual environment and text instructions. * Multilingual Support: to serve global users, besides English and Chinese, Qwen2-VL now supports the understanding of texts in different languages inside images, including most European languages, Japanese, Korean, Arabic, Vietnamese, etc. #### Model Architecture Updates: * Naive Dynamic Resolution: Unlike before, Qwen2-VL can handle arbitrary image resolutions, mapping them into a dynamic number of visual tokens, offering a more human-like visual processing experience. <p align="center"> <img src="https://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/qwen2_vl.jpg" width="80%"/> <p> * Multimodal Rotary Position Embedding (M-ROPE): Decomposes positional embedding into parts to capture 1D textual, 2D visual, and 3D video positional information, enhancing its multimodal processing capabilities. <p align="center"> <img src="http://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/mrope.png" width="80%"/> <p> We have three models with 2, 7 and 72 billion parameters. This repo contains the instruction-tuned 2B Qwen2-VL model. For more information, visit our [Blog](https://qwenlm.github.io/blog/qwen2-vl/) and [GitHub](https://github.com/QwenLM/Qwen2-VL). ## Evaluation ### Image Benchmarks \| Benchmark \| InternVL2-2B \| MiniCPM-V 2.0 \| Qwen2-VL-2B \| \| :--- \| :---: \| :---: \| :---: \| \| MMMU<sub>val</sub> \| 36.3 \| 38.2 \| 41.1 \| \| DocVQA<sub>test</sub> \| 86.9 \| - \| 90.1 \| \| InfoVQA<sub>test</sub> \| 58.9 \| - \| 65.5 \| \| ChartQA<sub>test</sub> \| 76.2 \| - \| 73.5 \| \| TextVQA<sub>val</sub> \| 73.4 \| - \| 79.7 \| \| OCRBench \| 781 \| 605 \| 794 \| \| MTVQA \| - \| - \| 20.0 \| \| VCR<sub>en easy</sub> \| - \| - \| 81.45 \| VCR<sub>zh easy</sub> \| - \| - \| 46.16 \| RealWorldQA \| 57.3 \| 55.8 \| 62.9 \| \| MME<sub>sum</sub> \| 1876.8 \| 1808.6 \| 1872.0 \| \| MMBench-EN<sub>test</sub> \| 73.2 \| 69.1 \| 74.9 \| \| MMBench-CN<sub>test</sub> \| 70.9 \| 66.5 \| 73.5 \| \| MMBench-V1.1<sub>test</sub> \| 69.6 \| 65.8 \| 72.2 \| \| MMT-Bench<sub>test</sub> \| - \| - \| 54.5 \| \| MMStar \| 49.8 \| 39.1 \| 48.0 \| \| MMVet<sub>GPT-4-Turbo</sub> \| 39.7 \| 41.0 \| 49.5 \| \| HallBench<sub>avg</sub> \| 38.0 \| 36.1 \| 41.7 \| \| MathVista<sub>testmini</sub> \| 46.0 \| 39.8 \| 43.0 \| \| MathVision \| - \| - \| 12.4 \| ### Video Benchmarks \| Benchmark \| Qwen2-VL-2B \| \| :--- \| :---: \| \| MVBench \| 63.2 \| \| PerceptionTest<sub>test</sub> \| 53.9 \| \| EgoSchema<sub>test</sub> \| 54.9 \| \| Video-MME<sub>wo/w subs</sub> \| 55.6/60.4 \| ## Requirements The code of Qwen2-VL has been in the latest Hugging face transformers and we advise you to build from source with command `pip install git+https://github.com/huggingface/transformers`, or you might encounter the following error: ``` KeyError: 'qwen2_vl' ``` ## Quickstart We offer a toolkit to help you handle various types of visual input more conveniently. This includes base64, URLs, and interleaved images and videos. You can install it using the following command: ```bash pip install qwen-vl-utils ``` Here we show a code snippet to show you how to use the chat model with `transformers` and `qwen_vl_utils`: ```python from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor from qwen_vl_utils import process_vision_info # default: Load the model on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) # We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios. # model = Qwen2VLForConditionalGeneration.from_pretrained( # "Qwen/Qwen2-VL-2B-Instruct", # torch_dtype=torch.bfloat16, # attn_implementation="flash_attention_2", # device_map="auto", # ) # default processer processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage. # min_pixels = 2562828 # max_pixels = 12802828 # processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels) messages = [ { "role": "user", "content": [ { "type": "image", "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference: Generation of the output generated_ids = model.generate(inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` <details> <summary>Without qwen_vl_utils</summary> ```python from PIL import Image import requests import torch from torchvision import io from typing import Dict from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor # Load the model in half-precision on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # Image url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg" image = Image.open(requests.get(url, stream=True).raw) conversation = [ { "role": "user", "content": [ { "type": "image", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preprocess the inputs text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True) # Excepted output: '<\|im_start\|>system\nYou are a helpful assistant.<\|im_end\|>\n<\|im_start\|>user\n<\|vision_start\|><\|image_pad\|><\|vision_end\|>Describe this image.<\|im_end\|>\n<\|im_start\|>assistant\n' inputs = processor( text=[text_prompt], images=[image], padding=True, return_tensors="pt" ) inputs = inputs.to("cuda") # Inference: Generation of the output output_ids = model.generate(inputs, max_new_tokens=128) generated_ids = [ output_ids[len(input_ids) :] for input_ids, output_ids in zip(inputs.input_ids, output_ids) ] output_text = processor.batch_decode( generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True ) print(output_text) ``` </details> <details> <summary>Multi image inference</summary> ```python # Messages containing multiple images and a text query messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "Identify the similarities between these images."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(*inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Video inference</summary> ```python # Messages containing a images list as a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": [ "file:///path/to/frame1.jpg", "file:///path/to/frame2.jpg", "file:///path/to/frame3.jpg", "file:///path/to/frame4.jpg", ], "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Messages containing a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": "file:///path/to/video1.mp4", "max_pixels": 360 420, "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Batch inference</summary> ```python # Sample messages for batch inference messages1 = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "What are the common elements in these pictures?"}, ], } ] messages2 = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Who are you?"}, ] # Combine messages for batch processing messages = [messages1, messages1] # Preparation for batch inference texts = [ processor.apply_chat_template(msg, tokenize=False, add_generation_prompt=True) for msg in messages ] image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=texts, images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Batch Inference generated_ids = model.generate(inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_texts = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_texts) ``` </details> ### More Usage Tips For input images, we support local files, base64, and URLs. For videos, we currently only support local files. ```python # You can directly insert a local file path, a URL, or a base64-encoded image into the position where you want in the text. ## Local file path messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Image URL messages = [ { "role": "user", "content": [ {"type": "image", "image": "http://path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Base64 encoded image messages = [ { "role": "user", "content": [ {"type": "image", "image": "data:image;base64,/9j/..."}, {"type": "text", "text": "Describe this image."}, ], } ] ``` #### Image Resolution for performance boost The model supports a wide range of resolution inputs. By default, it uses the native resolution for input, but higher resolutions can enhance performance at the cost of more computation. Users can set the minimum and maximum number of pixels to achieve an optimal configuration for their needs, such as a token count range of 256-1280, to balance speed and memory usage. ```python min_pixels = 256 * 28 * 28 max_pixels = 1280 * 28 * 28 processor = AutoProcessor.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels ) ``` Besides, We provide two methods for fine-grained control over the image size input to the model: 1. Define min_pixels and max_pixels: Images will be resized to maintain their aspect ratio within the range of min_pixels and max_pixels. 2. Specify exact dimensions: Directly set `resized_height` and `resized_width`. These values will be rounded to the nearest multiple of 28. ```python # min_pixels and max_pixels messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "resized_height": 280, "resized_width": 420, }, {"type": "text", "text": "Describe this image."}, ], } ] # resized_height and resized_width messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "min_pixels": 50176, "max_pixels": 50176, }, {"type": "text", "text": "Describe this image."}, ], } ] ``` ## Limitations While Qwen2-VL are applicable to a wide range of visual tasks, it is equally important to understand its limitations. Here are some known restrictions: 1. Lack of Audio Support: The current model does not comprehend audio information within videos. 2. Data timeliness: Our image dataset is updated until June 2023, and information subsequent to this date may not be covered. 3. Constraints in Individuals and Intellectual Property (IP): The model's capacity to recognize specific individuals or IPs is limited, potentially failing to comprehensively cover all well-known personalities or brands. 4. Limited Capacity for Complex Instruction: When faced with intricate multi-step instructions, the model's understanding and execution capabilities require enhancement. 5. Insufficient Counting Accuracy: Particularly in complex scenes, the accuracy of object counting is not high, necessitating further improvements. 6. Weak Spatial Reasoning Skills: Especially in 3D spaces, the model's inference of object positional relationships is inadequate, making it difficult to precisely judge the relative positions of objects. These limitations serve as ongoing directions for model optimization and improvement, and we are committed to continually enhancing the model's performance and scope of application. ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{Qwen2VL, title={Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution}, author={Wang, Peng and Bai, Shuai and Tan, Sinan and Wang, Shijie and Fan, Zhihao and Bai, Jinze and Chen, Keqin and Liu, Xuejing and Wang, Jialin and Ge, Wenbin and Fan, Yang and Dang, Kai and Du, Mengfei and Ren, Xuancheng and Men, Rui and Liu, Dayiheng and Zhou, Chang and Zhou, Jingren and Lin, Junyang}, journal={arXiv preprint arXiv:2409.12191}, year={2024} } @article{Qwen-VL, title={Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond}, author={Bai, Jinze and Bai, Shuai and Yang, Shusheng and Wang, Shijie and Tan, Sinan and Wang, Peng and Lin, Junyang and Zhou, Chang and Zhou, Jingren}, journal={arXiv preprint arXiv:2308.12966}, year={2023} } ```	{"base_model": ["Qwen/Qwen2-VL-2B"], "language": ["en"], "library_name": "transformers", "license": "apache-2.0", "pipeline_tag": "image-text-to-text", "tags": ["multimodal"]}	task	[ "QUESTION_ANSWERING" ]	44,616
Netta1994/setfit_baai_2k_fixed	Netta1994	text-classification	[ "sentence-transformers", "safetensors", "bert", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	2024-05-30T13:59:14Z	2024-05-30T13:59:48+00:00	4	0	--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- # Netta1994/setfit_baai_2k_fixed This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Netta1994/setfit_baai_2k_fixed") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	null	Non_BioNLP	# Netta1994/setfit_baai_2k_fixed This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Netta1994/setfit_baai_2k_fixed") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}	task	[ "TEXT_CLASSIFICATION" ]	44,617
gokulsrinivasagan/distilbert_lda_5_mrpc	gokulsrinivasagan	text-classification	[ "transformers", "tensorboard", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "en", "dataset:glue", "base_model:gokulsrinivasagan/distilbert_lda_5", "base_model:finetune:gokulsrinivasagan/distilbert_lda_5", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2024-11-22T09:17:36Z	2024-11-22T09:19:21+00:00	16	0	--- base_model: gokulsrinivasagan/distilbert_lda_5 datasets: - glue language: - en library_name: transformers metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: distilbert_lda_5_mrpc results: - task: type: text-classification name: Text Classification dataset: name: GLUE MRPC type: glue args: mrpc metrics: - type: accuracy value: 0.6838235294117647 name: Accuracy - type: f1 value: 0.8122270742358079 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert_lda_5_mrpc This model is a fine-tuned version of [gokulsrinivasagan/distilbert_lda_5](https://huggingface.co/gokulsrinivasagan/distilbert_lda_5) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.6231 - Accuracy: 0.6838 - F1: 0.8122 - Combined Score: 0.7480 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 30 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| Combined Score \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\|:--------------:\| \| 0.8361 \| 1.0 \| 15 \| 0.6269 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6378 \| 2.0 \| 30 \| 0.6249 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6312 \| 3.0 \| 45 \| 0.6317 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6376 \| 4.0 \| 60 \| 0.6263 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.635 \| 5.0 \| 75 \| 0.6289 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6344 \| 6.0 \| 90 \| 0.6243 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6332 \| 7.0 \| 105 \| 0.6244 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.634 \| 8.0 \| 120 \| 0.6235 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6295 \| 9.0 \| 135 \| 0.6311 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6359 \| 10.0 \| 150 \| 0.6260 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6356 \| 11.0 \| 165 \| 0.6248 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6329 \| 12.0 \| 180 \| 0.6231 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6311 \| 13.0 \| 195 \| 0.6242 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6348 \| 14.0 \| 210 \| 0.6261 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6329 \| 15.0 \| 225 \| 0.6254 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6306 \| 16.0 \| 240 \| 0.6252 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6296 \| 17.0 \| 255 \| 0.6247 \| 0.6838 \| 0.8122 \| 0.7480 \| ### Framework versions - Transformers 4.46.3 - Pytorch 2.2.1+cu118 - Datasets 2.17.0 - Tokenizers 0.20.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert_lda_5_mrpc This model is a fine-tuned version of [gokulsrinivasagan/distilbert_lda_5](https://huggingface.co/gokulsrinivasagan/distilbert_lda_5) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.6231 - Accuracy: 0.6838 - F1: 0.8122 - Combined Score: 0.7480 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 30 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| Combined Score \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\|:--------------:\| \| 0.8361 \| 1.0 \| 15 \| 0.6269 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6378 \| 2.0 \| 30 \| 0.6249 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6312 \| 3.0 \| 45 \| 0.6317 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6376 \| 4.0 \| 60 \| 0.6263 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.635 \| 5.0 \| 75 \| 0.6289 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6344 \| 6.0 \| 90 \| 0.6243 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6332 \| 7.0 \| 105 \| 0.6244 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.634 \| 8.0 \| 120 \| 0.6235 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6295 \| 9.0 \| 135 \| 0.6311 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6359 \| 10.0 \| 150 \| 0.6260 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6356 \| 11.0 \| 165 \| 0.6248 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6329 \| 12.0 \| 180 \| 0.6231 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6311 \| 13.0 \| 195 \| 0.6242 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6348 \| 14.0 \| 210 \| 0.6261 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6329 \| 15.0 \| 225 \| 0.6254 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6306 \| 16.0 \| 240 \| 0.6252 \| 0.6838 \| 0.8122 \| 0.7480 \| \| 0.6296 \| 17.0 \| 255 \| 0.6247 \| 0.6838 \| 0.8122 \| 0.7480 \| ### Framework versions - Transformers 4.46.3 - Pytorch 2.2.1+cu118 - Datasets 2.17.0 - Tokenizers 0.20.3	{"base_model": "gokulsrinivasagan/distilbert_lda_5", "datasets": ["glue"], "language": ["en"], "library_name": "transformers", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert_lda_5_mrpc", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "GLUE MRPC", "type": "glue", "args": "mrpc"}, "metrics": [{"type": "accuracy", "value": 0.6838235294117647, "name": "Accuracy"}, {"type": "f1", "value": 0.8122270742358079, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,618
SeungWonSeo/bert-base-cased	SeungWonSeo	null	[ "pytorch", "tf", "jax", "safetensors", "bert", "exbert", "en", "dataset:bookcorpus", "dataset:wikipedia", "arxiv:1810.04805", "license:apache-2.0", "region:us" ]	2025-03-11T13:51:06Z	2025-03-11T13:58:38+00:00	149	0	--- datasets: - bookcorpus - wikipedia language: en license: apache-2.0 tags: - exbert --- # BERT base model (cased) Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in [this paper](https://arxiv.org/abs/1810.04805) and first released in [this repository](https://github.com/google-research/bert). This model is case-sensitive: it makes a difference between english and English. Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by the Hugging Face team. ## Model description BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was pretrained with two objectives: - Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run the entire masked sentence through the model and has to predict the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the sentence. - Next sentence prediction (NSP): the models concatenates two masked sentences as inputs during pretraining. Sometimes they correspond to sentences that were next to each other in the original text, sometimes not. The model then has to predict if the two sentences were following each other or not. This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard classifier using the features produced by the BERT model as inputs. ## Intended uses & limitations You can use the raw model for either masked language modeling or next sentence prediction, but it's mostly intended to be fine-tuned on a downstream task. See the [model hub](https://huggingface.co/models?filter=bert) to look for fine-tuned versions on a task that interests you. Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked) to make decisions, such as sequence classification, token classification or question answering. For tasks such as text generation you should look at model like GPT2. ### How to use You can use this model directly with a pipeline for masked language modeling: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("Hello I'm a [MASK] model.") [{'sequence': "[CLS] Hello I'm a fashion model. [SEP]", 'score': 0.09019174426794052, 'token': 4633, 'token_str': 'fashion'}, {'sequence': "[CLS] Hello I'm a new model. [SEP]", 'score': 0.06349995732307434, 'token': 1207, 'token_str': 'new'}, {'sequence': "[CLS] Hello I'm a male model. [SEP]", 'score': 0.06228214129805565, 'token': 2581, 'token_str': 'male'}, {'sequence': "[CLS] Hello I'm a professional model. [SEP]", 'score': 0.0441727414727211, 'token': 1848, 'token_str': 'professional'}, {'sequence': "[CLS] Hello I'm a super model. [SEP]", 'score': 0.03326151892542839, 'token': 7688, 'token_str': 'super'}] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = BertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = TFBertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ### Limitations and bias Even if the training data used for this model could be characterized as fairly neutral, this model can have biased predictions: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("The man worked as a [MASK].") [{'sequence': '[CLS] The man worked as a lawyer. [SEP]', 'score': 0.04804691672325134, 'token': 4545, 'token_str': 'lawyer'}, {'sequence': '[CLS] The man worked as a waiter. [SEP]', 'score': 0.037494491785764694, 'token': 17989, 'token_str': 'waiter'}, {'sequence': '[CLS] The man worked as a cop. [SEP]', 'score': 0.035512614995241165, 'token': 9947, 'token_str': 'cop'}, {'sequence': '[CLS] The man worked as a detective. [SEP]', 'score': 0.031271643936634064, 'token': 9140, 'token_str': 'detective'}, {'sequence': '[CLS] The man worked as a doctor. [SEP]', 'score': 0.027423162013292313, 'token': 3995, 'token_str': 'doctor'}] >>> unmasker("The woman worked as a [MASK].") [{'sequence': '[CLS] The woman worked as a nurse. [SEP]', 'score': 0.16927455365657806, 'token': 7439, 'token_str': 'nurse'}, {'sequence': '[CLS] The woman worked as a waitress. [SEP]', 'score': 0.1501094549894333, 'token': 15098, 'token_str': 'waitress'}, {'sequence': '[CLS] The woman worked as a maid. [SEP]', 'score': 0.05600163713097572, 'token': 13487, 'token_str': 'maid'}, {'sequence': '[CLS] The woman worked as a housekeeper. [SEP]', 'score': 0.04838843643665314, 'token': 26458, 'token_str': 'housekeeper'}, {'sequence': '[CLS] The woman worked as a cook. [SEP]', 'score': 0.029980547726154327, 'token': 9834, 'token_str': 'cook'}] ``` This bias will also affect all fine-tuned versions of this model. ## Training data The BERT model was pretrained on [BookCorpus](https://yknzhu.wixsite.com/mbweb), a dataset consisting of 11,038 unpublished books and [English Wikipedia](https://en.wikipedia.org/wiki/English_Wikipedia) (excluding lists, tables and headers). ## Training procedure ### Preprocessing The texts are tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are then of the form: ``` [CLS] Sentence A [SEP] Sentence B [SEP] ``` With probability 0.5, sentence A and sentence B correspond to two consecutive sentences in the original corpus and in the other cases, it's another random sentence in the corpus. Note that what is considered a sentence here is a consecutive span of text usually longer than a single sentence. The only constrain is that the result with the two "sentences" has a combined length of less than 512 tokens. The details of the masking procedure for each sentence are the following: - 15% of the tokens are masked. - In 80% of the cases, the masked tokens are replaced by `[MASK]`. - In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace. - In the 10% remaining cases, the masked tokens are left as is. ### Pretraining The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total) for one million steps with a batch size of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer used is Adam with a learning rate of 1e-4, \\(\beta_{1} = 0.9\\) and \\(\beta_{2} = 0.999\\), a weight decay of 0.01, learning rate warmup for 10,000 steps and linear decay of the learning rate after. ## Evaluation results When fine-tuned on downstream tasks, this model achieves the following results: Glue test results: \| Task \| MNLI-(m/mm) \| QQP \| QNLI \| SST-2 \| CoLA \| STS-B \| MRPC \| RTE \| Average \| \|:----:\|:-----------:\|:----:\|:----:\|:-----:\|:----:\|:-----:\|:----:\|:----:\|:-------:\| \| \| 84.6/83.4 \| 71.2 \| 90.5 \| 93.5 \| 52.1 \| 85.8 \| 88.9 \| 66.4 \| 79.6 \| ### BibTeX entry and citation info ```bibtex @article{DBLP:journals/corr/abs-1810-04805, author = {Jacob Devlin and Ming{-}Wei Chang and Kenton Lee and Kristina Toutanova}, title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language Understanding}, journal = {CoRR}, volume = {abs/1810.04805}, year = {2018}, url = {http://arxiv.org/abs/1810.04805}, archivePrefix = {arXiv}, eprint = {1810.04805}, timestamp = {Tue, 30 Oct 2018 20:39:56 +0100}, biburl = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ``` <a href="https://huggingface.co/exbert/?model=bert-base-cased"> <img width="300px" src="https://cdn-media.huggingface.co/exbert/button.png"> </a>	null	Non_BioNLP	# BERT base model (cased) Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in [this paper](https://arxiv.org/abs/1810.04805) and first released in [this repository](https://github.com/google-research/bert). This model is case-sensitive: it makes a difference between english and English. Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by the Hugging Face team. ## Model description BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was pretrained with two objectives: - Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run the entire masked sentence through the model and has to predict the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the sentence. - Next sentence prediction (NSP): the models concatenates two masked sentences as inputs during pretraining. Sometimes they correspond to sentences that were next to each other in the original text, sometimes not. The model then has to predict if the two sentences were following each other or not. This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard classifier using the features produced by the BERT model as inputs. ## Intended uses & limitations You can use the raw model for either masked language modeling or next sentence prediction, but it's mostly intended to be fine-tuned on a downstream task. See the [model hub](https://huggingface.co/models?filter=bert) to look for fine-tuned versions on a task that interests you. Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked) to make decisions, such as sequence classification, token classification or question answering. For tasks such as text generation you should look at model like GPT2. ### How to use You can use this model directly with a pipeline for masked language modeling: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("Hello I'm a [MASK] model.") [{'sequence': "[CLS] Hello I'm a fashion model. [SEP]", 'score': 0.09019174426794052, 'token': 4633, 'token_str': 'fashion'}, {'sequence': "[CLS] Hello I'm a new model. [SEP]", 'score': 0.06349995732307434, 'token': 1207, 'token_str': 'new'}, {'sequence': "[CLS] Hello I'm a male model. [SEP]", 'score': 0.06228214129805565, 'token': 2581, 'token_str': 'male'}, {'sequence': "[CLS] Hello I'm a professional model. [SEP]", 'score': 0.0441727414727211, 'token': 1848, 'token_str': 'professional'}, {'sequence': "[CLS] Hello I'm a super model. [SEP]", 'score': 0.03326151892542839, 'token': 7688, 'token_str': 'super'}] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = BertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = TFBertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ### Limitations and bias Even if the training data used for this model could be characterized as fairly neutral, this model can have biased predictions: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("The man worked as a [MASK].") [{'sequence': '[CLS] The man worked as a lawyer. [SEP]', 'score': 0.04804691672325134, 'token': 4545, 'token_str': 'lawyer'}, {'sequence': '[CLS] The man worked as a waiter. [SEP]', 'score': 0.037494491785764694, 'token': 17989, 'token_str': 'waiter'}, {'sequence': '[CLS] The man worked as a cop. [SEP]', 'score': 0.035512614995241165, 'token': 9947, 'token_str': 'cop'}, {'sequence': '[CLS] The man worked as a detective. [SEP]', 'score': 0.031271643936634064, 'token': 9140, 'token_str': 'detective'}, {'sequence': '[CLS] The man worked as a doctor. [SEP]', 'score': 0.027423162013292313, 'token': 3995, 'token_str': 'doctor'}] >>> unmasker("The woman worked as a [MASK].") [{'sequence': '[CLS] The woman worked as a nurse. [SEP]', 'score': 0.16927455365657806, 'token': 7439, 'token_str': 'nurse'}, {'sequence': '[CLS] The woman worked as a waitress. [SEP]', 'score': 0.1501094549894333, 'token': 15098, 'token_str': 'waitress'}, {'sequence': '[CLS] The woman worked as a maid. [SEP]', 'score': 0.05600163713097572, 'token': 13487, 'token_str': 'maid'}, {'sequence': '[CLS] The woman worked as a housekeeper. [SEP]', 'score': 0.04838843643665314, 'token': 26458, 'token_str': 'housekeeper'}, {'sequence': '[CLS] The woman worked as a cook. [SEP]', 'score': 0.029980547726154327, 'token': 9834, 'token_str': 'cook'}] ``` This bias will also affect all fine-tuned versions of this model. ## Training data The BERT model was pretrained on [BookCorpus](https://yknzhu.wixsite.com/mbweb), a dataset consisting of 11,038 unpublished books and [English Wikipedia](https://en.wikipedia.org/wiki/English_Wikipedia) (excluding lists, tables and headers). ## Training procedure ### Preprocessing The texts are tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are then of the form: ``` [CLS] Sentence A [SEP] Sentence B [SEP] ``` With probability 0.5, sentence A and sentence B correspond to two consecutive sentences in the original corpus and in the other cases, it's another random sentence in the corpus. Note that what is considered a sentence here is a consecutive span of text usually longer than a single sentence. The only constrain is that the result with the two "sentences" has a combined length of less than 512 tokens. The details of the masking procedure for each sentence are the following: - 15% of the tokens are masked. - In 80% of the cases, the masked tokens are replaced by `[MASK]`. - In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace. - In the 10% remaining cases, the masked tokens are left as is. ### Pretraining The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total) for one million steps with a batch size of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer used is Adam with a learning rate of 1e-4, \\(\beta_{1} = 0.9\\) and \\(\beta_{2} = 0.999\\), a weight decay of 0.01, learning rate warmup for 10,000 steps and linear decay of the learning rate after. ## Evaluation results When fine-tuned on downstream tasks, this model achieves the following results: Glue test results: \| Task \| MNLI-(m/mm) \| QQP \| QNLI \| SST-2 \| CoLA \| STS-B \| MRPC \| RTE \| Average \| \|:----:\|:-----------:\|:----:\|:----:\|:-----:\|:----:\|:-----:\|:----:\|:----:\|:-------:\| \| \| 84.6/83.4 \| 71.2 \| 90.5 \| 93.5 \| 52.1 \| 85.8 \| 88.9 \| 66.4 \| 79.6 \| ### BibTeX entry and citation info ```bibtex @article{DBLP:journals/corr/abs-1810-04805, author = {Jacob Devlin and Ming{-}Wei Chang and Kenton Lee and Kristina Toutanova}, title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language Understanding}, journal = {CoRR}, volume = {abs/1810.04805}, year = {2018}, url = {http://arxiv.org/abs/1810.04805}, archivePrefix = {arXiv}, eprint = {1810.04805}, timestamp = {Tue, 30 Oct 2018 20:39:56 +0100}, biburl = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ``` <a href="https://huggingface.co/exbert/?model=bert-base-cased"> <img width="300px" src="https://cdn-media.huggingface.co/exbert/button.png"> </a>	{"datasets": ["bookcorpus", "wikipedia"], "language": "en", "license": "apache-2.0", "tags": ["exbert"]}	task	[ "QUESTION_ANSWERING" ]	44,619
Lots-of-LoRAs/Mistral-7B-Instruct-v0.2-4b-r16-task1121	Lots-of-LoRAs	null	[ "pytorch", "safetensors", "en", "arxiv:1910.09700", "arxiv:2407.00066", "base_model:mistralai/Mistral-7B-Instruct-v0.2", "base_model:finetune:mistralai/Mistral-7B-Instruct-v0.2", "license:mit", "region:us" ]	2025-01-03T18:30:17Z	2025-01-03T18:30:23+00:00	0	0	--- base_model: mistralai/Mistral-7B-Instruct-v0.2 language: en library_name: pytorch license: mit --- # Model Card for Mistral-7B-Instruct-v0.2-4b-r16-task1121 <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> LoRA trained on task1121_alt_ja_khm_translation - Developed by: bruel - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: LoRA - Language(s) (NLP): en - License: mit - Finetuned from model [optional]: mistralai/Mistral-7B-Instruct-v0.2 ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: https://github.com/bruel-gabrielsson - Paper [optional]: "Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead" (2024), Rickard Brüel Gabrielsson, Jiacheng Zhu, Onkar Bhardwaj, Leshem Choshen, Kristjan Greenewald, Mikhail Yurochkin and Justin Solomon - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> https://huggingface.co/datasets/Lots-of-LoRAs/task1121_alt_ja_khm_translation sourced from https://github.com/allenai/natural-instructions ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: @misc{brüelgabrielsson2024compressserveservingthousands, title={Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead}, author={Rickard Brüel-Gabrielsson and Jiacheng Zhu and Onkar Bhardwaj and Leshem Choshen and Kristjan Greenewald and Mikhail Yurochkin and Justin Solomon}, year={2024}, eprint={2407.00066}, archivePrefix={arXiv}, primaryClass={cs.DC}, url={https://arxiv.org/abs/2407.00066}, } APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	null	Non_BioNLP	# Model Card for Mistral-7B-Instruct-v0.2-4b-r16-task1121 <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> LoRA trained on task1121_alt_ja_khm_translation - Developed by: bruel - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: LoRA - Language(s) (NLP): en - License: mit - Finetuned from model [optional]: mistralai/Mistral-7B-Instruct-v0.2 ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: https://github.com/bruel-gabrielsson - Paper [optional]: "Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead" (2024), Rickard Brüel Gabrielsson, Jiacheng Zhu, Onkar Bhardwaj, Leshem Choshen, Kristjan Greenewald, Mikhail Yurochkin and Justin Solomon - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> https://huggingface.co/datasets/Lots-of-LoRAs/task1121_alt_ja_khm_translation sourced from https://github.com/allenai/natural-instructions ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: @misc{brüelgabrielsson2024compressserveservingthousands, title={Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead}, author={Rickard Brüel-Gabrielsson and Jiacheng Zhu and Onkar Bhardwaj and Leshem Choshen and Kristjan Greenewald and Mikhail Yurochkin and Justin Solomon}, year={2024}, eprint={2407.00066}, archivePrefix={arXiv}, primaryClass={cs.DC}, url={https://arxiv.org/abs/2407.00066}, } APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"base_model": "mistralai/Mistral-7B-Instruct-v0.2", "language": "en", "library_name": "pytorch", "license": "mit"}	task	[ "TRANSLATION" ]	44,620
dz5035/bicleaner-ai-full-large-de-xx	dz5035	null	[ "transformers", "tf", "xlm-roberta", "bicleaner-ai", "ar", "bn", "da", "de", "en", "es", "fa", "hi", "ja", "ko", "ru", "ta", "th", "tr", "uk", "vi", "zh", "multilingual", "license:cc-by-sa-4.0", "endpoints_compatible", "region:us" ]	2024-06-03T13:20:12Z	2024-06-03T13:27:34+00:00	6	0	--- language: - ar - bn - da - de - en - es - fa - hi - ja - ko - ru - ta - th - tr - uk - vi - zh - multilingual license: cc-by-sa-4.0 tags: - bicleaner-ai tasks: - text-classification --- # Bicleaner AI multilingual full large model for de-xx Bicleaner AI is a tool that aims at detecting noisy sentence pairs in a parallel corpus. It indicates the likelihood of a pair of sentences being mutual translations (with a value near to 1) or not (with a value near to 0). Sentence pairs considered very noisy are scored with 0. Find out at our repository for further instructions on how to use it: https://github.com/bitextor/bicleaner-ai	null	Non_BioNLP	# Bicleaner AI multilingual full large model for de-xx Bicleaner AI is a tool that aims at detecting noisy sentence pairs in a parallel corpus. It indicates the likelihood of a pair of sentences being mutual translations (with a value near to 1) or not (with a value near to 0). Sentence pairs considered very noisy are scored with 0. Find out at our repository for further instructions on how to use it: https://github.com/bitextor/bicleaner-ai	{"language": ["ar", "bn", "da", "de", "en", "es", "fa", "hi", "ja", "ko", "ru", "ta", "th", "tr", "uk", "vi", "zh", "multilingual"], "license": "cc-by-sa-4.0", "tags": ["bicleaner-ai"], "tasks": ["text-classification"]}	task	[ "TRANSLATION" ]	44,621
lujain5666/finetune-gemma-7b	lujain5666	sentence-similarity	[ "sentence-transformers", "safetensors", "bert", "feature-extraction", "sentence-similarity", "transformers", "unsloth", "en", "dataset:s2orc", "dataset:flax-sentence-embeddings/stackexchange_xml", "dataset:ms_marco", "dataset:gooaq", "dataset:yahoo_answers_topics", "dataset:code_search_net", "dataset:search_qa", "dataset:eli5", "dataset:snli", "dataset:multi_nli", "dataset:wikihow", "dataset:natural_questions", "dataset:trivia_qa", "dataset:embedding-data/sentence-compression", "dataset:embedding-data/flickr30k-captions", "dataset:embedding-data/altlex", "dataset:embedding-data/simple-wiki", "dataset:embedding-data/QQP", "dataset:embedding-data/SPECTER", "dataset:embedding-data/PAQ_pairs", "dataset:embedding-data/WikiAnswers", "arxiv:1904.06472", "arxiv:2102.07033", "arxiv:2104.08727", "arxiv:1704.05179", "arxiv:1810.09305", "license:apache-2.0", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	2024-08-21T18:32:45Z	2024-08-22T14:49:25+00:00	9	0	--- datasets: - s2orc - flax-sentence-embeddings/stackexchange_xml - ms_marco - gooaq - yahoo_answers_topics - code_search_net - search_qa - eli5 - snli - multi_nli - wikihow - natural_questions - trivia_qa - embedding-data/sentence-compression - embedding-data/flickr30k-captions - embedding-data/altlex - embedding-data/simple-wiki - embedding-data/QQP - embedding-data/SPECTER - embedding-data/PAQ_pairs - embedding-data/WikiAnswers language: en library_name: sentence-transformers license: apache-2.0 pipeline_tag: sentence-similarity tags: - sentence-transformers - feature-extraction - sentence-similarity - transformers - unsloth --- # all-MiniLM-L6-v2 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer sentences = ["This is an example sentence", "Each sentence is converted"] model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2') embeddings = model.encode(sentences) print(embeddings) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch import torch.nn.functional as F #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) # Sentences we want sentence embeddings for sentences = ['This is an example sentence', 'Each sentence is converted'] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') model = AutoModel.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') # Tokenize sentences encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(*encoded_input) # Perform pooling sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask']) # Normalize embeddings sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1) print("Sentence embeddings:") print(sentence_embeddings) ``` ## Evaluation Results For an automated evaluation of this model, see the Sentence Embeddings Benchmark: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/all-MiniLM-L6-v2) ------ ## Background The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model and fine-tuned in on a 1B sentence pairs dataset. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset. We developed this model during the [Community week using JAX/Flax for NLP & CV](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104), organized by Hugging Face. We developed this model as part of the project: [Train the Best Sentence Embedding Model Ever with 1B Training Pairs](https://discuss.huggingface.co/t/train-the-best-sentence-embedding-model-ever-with-1b-training-pairs/7354). We benefited from efficient hardware infrastructure to run the project: 7 TPUs v3-8, as well as intervention from Googles Flax, JAX, and Cloud team member about efficient deep learning frameworks. ## Intended uses Our model is intended to be used as a sentence and short paragraph encoder. Given an input text, it outputs a vector which captures the semantic information. The sentence vector may be used for information retrieval, clustering or sentence similarity tasks. By default, input text longer than 256 word pieces is truncated. ## Training procedure ### Pre-training We use the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model. Please refer to the model card for more detailed information about the pre-training procedure. ### Fine-tuning We fine-tune the model using a contrastive objective. Formally, we compute the cosine similarity from each possible sentence pairs from the batch. We then apply the cross entropy loss by comparing with true pairs. #### Hyper parameters We trained our model on a TPU v3-8. We train the model during 100k steps using a batch size of 1024 (128 per TPU core). We use a learning rate warm up of 500. The sequence length was limited to 128 tokens. We used the AdamW optimizer with a 2e-5 learning rate. The full training script is accessible in this current repository: `train_script.py`. #### Training data We use the concatenation from multiple datasets to fine-tune our model. The total number of sentence pairs is above 1 billion sentences. We sampled each dataset given a weighted probability which configuration is detailed in the `data_config.json` file. \| Dataset \| Paper \| Number of training tuples \| \|--------------------------------------------------------\|:----------------------------------------:\|:--------------------------:\| \| [Reddit comments (2015-2018)](https://github.com/PolyAI-LDN/conversational-datasets/tree/master/reddit) \| [paper](https://arxiv.org/abs/1904.06472) \| 726,484,430 \| \| [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Abstracts) \| [paper](https://aclanthology.org/2020.acl-main.447/) \| 116,288,806 \| \| [WikiAnswers](https://github.com/afader/oqa#wikianswers-corpus) Duplicate question pairs \| [paper](https://doi.org/10.1145/2623330.2623677) \| 77,427,422 \| \| [PAQ](https://github.com/facebookresearch/PAQ) (Question, Answer) pairs \| [paper](https://arxiv.org/abs/2102.07033) \| 64,371,441 \| \| [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Titles) \| [paper](https://aclanthology.org/2020.acl-main.447/) \| 52,603,982 \| \| [S2ORC](https://github.com/allenai/s2orc) (Title, Abstract) \| [paper](https://aclanthology.org/2020.acl-main.447/) \| 41,769,185 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Body) pairs \| - \| 25,316,456 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title+Body, Answer) pairs \| - \| 21,396,559 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Answer) pairs \| - \| 21,396,559 \| \| [MS MARCO](https://microsoft.github.io/msmarco/) triplets \| [paper](https://doi.org/10.1145/3404835.3462804) \| 9,144,553 \| \| [GOOAQ: Open Question Answering with Diverse Answer Types](https://github.com/allenai/gooaq) \| [paper](https://arxiv.org/pdf/2104.08727.pdf) \| 3,012,496 \| \| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Answer) \| [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) \| 1,198,260 \| \| [Code Search](https://huggingface.co/datasets/code_search_net) \| - \| 1,151,414 \| \| [COCO](https://cocodataset.org/#home) Image captions \| [paper](https://link.springer.com/chapter/10.1007%2F978-3-319-10602-1_48) \| 828,395\| \| [SPECTER](https://github.com/allenai/specter) citation triplets \| [paper](https://doi.org/10.18653/v1/2020.acl-main.207) \| 684,100 \| \| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Question, Answer) \| [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) \| 681,164 \| \| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Question) \| [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) \| 659,896 \| \| [SearchQA](https://huggingface.co/datasets/search_qa) \| [paper](https://arxiv.org/abs/1704.05179) \| 582,261 \| \| [Eli5](https://huggingface.co/datasets/eli5) \| [paper](https://doi.org/10.18653/v1/p19-1346) \| 325,475 \| \| [Flickr 30k](https://shannon.cs.illinois.edu/DenotationGraph/) \| [paper](https://transacl.org/ojs/index.php/tacl/article/view/229/33) \| 317,695 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles) \| \| 304,525 \| \| AllNLI ([SNLI](https://nlp.stanford.edu/projects/snli/) and [MultiNLI](https://cims.nyu.edu/~sbowman/multinli/) \| [paper SNLI](https://doi.org/10.18653/v1/d15-1075), [paper MultiNLI](https://doi.org/10.18653/v1/n18-1101) \| 277,230 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (bodies) \| \| 250,519 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles+bodies) \| \| 250,460 \| \| [Sentence Compression](https://github.com/google-research-datasets/sentence-compression) \| [paper](https://www.aclweb.org/anthology/D13-1155/) \| 180,000 \| \| [Wikihow](https://github.com/pvl/wikihow_pairs_dataset) \| [paper](https://arxiv.org/abs/1810.09305) \| 128,542 \| \| [Altlex](https://github.com/chridey/altlex/) \| [paper](https://aclanthology.org/P16-1135.pdf) \| 112,696 \| \| [Quora Question Triplets](https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs) \| - \| 103,663 \| \| [Simple Wikipedia](https://cs.pomona.edu/~dkauchak/simplification/) \| [paper](https://www.aclweb.org/anthology/P11-2117/) \| 102,225 \| \| [Natural Questions (NQ)](https://ai.google.com/research/NaturalQuestions) \| [paper](https://transacl.org/ojs/index.php/tacl/article/view/1455) \| 100,231 \| \| [SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) \| [paper](https://aclanthology.org/P18-2124.pdf) \| 87,599 \| \| [TriviaQA](https://huggingface.co/datasets/trivia_qa) \| - \| 73,346 \| \| Total* \| \| 1,170,060,424 \|	null	Non_BioNLP	# all-MiniLM-L6-v2 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer sentences = ["This is an example sentence", "Each sentence is converted"] model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2') embeddings = model.encode(sentences) print(embeddings) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch import torch.nn.functional as F #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) # Sentences we want sentence embeddings for sentences = ['This is an example sentence', 'Each sentence is converted'] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') model = AutoModel.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') # Tokenize sentences encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(*encoded_input) # Perform pooling sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask']) # Normalize embeddings sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1) print("Sentence embeddings:") print(sentence_embeddings) ``` ## Evaluation Results For an automated evaluation of this model, see the Sentence Embeddings Benchmark: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/all-MiniLM-L6-v2) ------ ## Background The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model and fine-tuned in on a 1B sentence pairs dataset. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset. We developed this model during the [Community week using JAX/Flax for NLP & CV](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104), organized by Hugging Face. We developed this model as part of the project: [Train the Best Sentence Embedding Model Ever with 1B Training Pairs](https://discuss.huggingface.co/t/train-the-best-sentence-embedding-model-ever-with-1b-training-pairs/7354). We benefited from efficient hardware infrastructure to run the project: 7 TPUs v3-8, as well as intervention from Googles Flax, JAX, and Cloud team member about efficient deep learning frameworks. ## Intended uses Our model is intended to be used as a sentence and short paragraph encoder. Given an input text, it outputs a vector which captures the semantic information. The sentence vector may be used for information retrieval, clustering or sentence similarity tasks. By default, input text longer than 256 word pieces is truncated. ## Training procedure ### Pre-training We use the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model. Please refer to the model card for more detailed information about the pre-training procedure. ### Fine-tuning We fine-tune the model using a contrastive objective. Formally, we compute the cosine similarity from each possible sentence pairs from the batch. We then apply the cross entropy loss by comparing with true pairs. #### Hyper parameters We trained our model on a TPU v3-8. We train the model during 100k steps using a batch size of 1024 (128 per TPU core). We use a learning rate warm up of 500. The sequence length was limited to 128 tokens. We used the AdamW optimizer with a 2e-5 learning rate. The full training script is accessible in this current repository: `train_script.py`. #### Training data We use the concatenation from multiple datasets to fine-tune our model. The total number of sentence pairs is above 1 billion sentences. We sampled each dataset given a weighted probability which configuration is detailed in the `data_config.json` file. \| Dataset \| Paper \| Number of training tuples \| \|--------------------------------------------------------\|:----------------------------------------:\|:--------------------------:\| \| [Reddit comments (2015-2018)](https://github.com/PolyAI-LDN/conversational-datasets/tree/master/reddit) \| [paper](https://arxiv.org/abs/1904.06472) \| 726,484,430 \| \| [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Abstracts) \| [paper](https://aclanthology.org/2020.acl-main.447/) \| 116,288,806 \| \| [WikiAnswers](https://github.com/afader/oqa#wikianswers-corpus) Duplicate question pairs \| [paper](https://doi.org/10.1145/2623330.2623677) \| 77,427,422 \| \| [PAQ](https://github.com/facebookresearch/PAQ) (Question, Answer) pairs \| [paper](https://arxiv.org/abs/2102.07033) \| 64,371,441 \| \| [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Titles) \| [paper](https://aclanthology.org/2020.acl-main.447/) \| 52,603,982 \| \| [S2ORC](https://github.com/allenai/s2orc) (Title, Abstract) \| [paper](https://aclanthology.org/2020.acl-main.447/) \| 41,769,185 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Body) pairs \| - \| 25,316,456 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title+Body, Answer) pairs \| - \| 21,396,559 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Answer) pairs \| - \| 21,396,559 \| \| [MS MARCO](https://microsoft.github.io/msmarco/) triplets \| [paper](https://doi.org/10.1145/3404835.3462804) \| 9,144,553 \| \| [GOOAQ: Open Question Answering with Diverse Answer Types](https://github.com/allenai/gooaq) \| [paper](https://arxiv.org/pdf/2104.08727.pdf) \| 3,012,496 \| \| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Answer) \| [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) \| 1,198,260 \| \| [Code Search](https://huggingface.co/datasets/code_search_net) \| - \| 1,151,414 \| \| [COCO](https://cocodataset.org/#home) Image captions \| [paper](https://link.springer.com/chapter/10.1007%2F978-3-319-10602-1_48) \| 828,395\| \| [SPECTER](https://github.com/allenai/specter) citation triplets \| [paper](https://doi.org/10.18653/v1/2020.acl-main.207) \| 684,100 \| \| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Question, Answer) \| [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) \| 681,164 \| \| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Question) \| [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) \| 659,896 \| \| [SearchQA](https://huggingface.co/datasets/search_qa) \| [paper](https://arxiv.org/abs/1704.05179) \| 582,261 \| \| [Eli5](https://huggingface.co/datasets/eli5) \| [paper](https://doi.org/10.18653/v1/p19-1346) \| 325,475 \| \| [Flickr 30k](https://shannon.cs.illinois.edu/DenotationGraph/) \| [paper](https://transacl.org/ojs/index.php/tacl/article/view/229/33) \| 317,695 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles) \| \| 304,525 \| \| AllNLI ([SNLI](https://nlp.stanford.edu/projects/snli/) and [MultiNLI](https://cims.nyu.edu/~sbowman/multinli/) \| [paper SNLI](https://doi.org/10.18653/v1/d15-1075), [paper MultiNLI](https://doi.org/10.18653/v1/n18-1101) \| 277,230 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (bodies) \| \| 250,519 \| \| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles+bodies) \| \| 250,460 \| \| [Sentence Compression](https://github.com/google-research-datasets/sentence-compression) \| [paper](https://www.aclweb.org/anthology/D13-1155/) \| 180,000 \| \| [Wikihow](https://github.com/pvl/wikihow_pairs_dataset) \| [paper](https://arxiv.org/abs/1810.09305) \| 128,542 \| \| [Altlex](https://github.com/chridey/altlex/) \| [paper](https://aclanthology.org/P16-1135.pdf) \| 112,696 \| \| [Quora Question Triplets](https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs) \| - \| 103,663 \| \| [Simple Wikipedia](https://cs.pomona.edu/~dkauchak/simplification/) \| [paper](https://www.aclweb.org/anthology/P11-2117/) \| 102,225 \| \| [Natural Questions (NQ)](https://ai.google.com/research/NaturalQuestions) \| [paper](https://transacl.org/ojs/index.php/tacl/article/view/1455) \| 100,231 \| \| [SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) \| [paper](https://aclanthology.org/P18-2124.pdf) \| 87,599 \| \| [TriviaQA](https://huggingface.co/datasets/trivia_qa) \| - \| 73,346 \| \| Total* \| \| 1,170,060,424 \|	{"datasets": ["s2orc", "flax-sentence-embeddings/stackexchange_xml", "ms_marco", "gooaq", "yahoo_answers_topics", "code_search_net", "search_qa", "eli5", "snli", "multi_nli", "wikihow", "natural_questions", "trivia_qa", "embedding-data/sentence-compression", "embedding-data/flickr30k-captions", "embedding-data/altlex", "embedding-data/simple-wiki", "embedding-data/QQP", "embedding-data/SPECTER", "embedding-data/PAQ_pairs", "embedding-data/WikiAnswers"], "language": "en", "library_name": "sentence-transformers", "license": "apache-2.0", "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "feature-extraction", "sentence-similarity", "transformers", "unsloth"]}	task	[ "QUESTION_ANSWERING" ]	44,622
amitesh11/bart-finance-pegasus	amitesh11	summarization	[ "transformers", "pytorch", "bart", "text2text-generation", "summarization", "en", "license:ms-pl", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-10-17T11:14:14Z	2023-10-17T11:32:10+00:00	111	1	--- language: - en license: ms-pl pipeline_tag: summarization ---	null	Non_BioNLP		{"language": ["en"], "license": "ms-pl", "pipeline_tag": "summarization"}	task	[ "SUMMARIZATION" ]	44,623
UBC-NLP/AraT5-base-title-generation	UBC-NLP	text2text-generation	[ "transformers", "pytorch", "tf", "t5", "text2text-generation", "Arabic T5", "MSA", "Twitter", "Arabic Dialect", "Arabic Machine Translation", "Arabic Text Summarization", "Arabic News Title and Question Generation", "Arabic Paraphrasing and Transliteration", "Arabic Code-Switched Translation", "ar", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:05Z	2022-05-26T18:29:45+00:00	50	12	--- language: - ar tags: - Arabic T5 - MSA - Twitter - Arabic Dialect - Arabic Machine Translation - Arabic Text Summarization - Arabic News Title and Question Generation - Arabic Paraphrasing and Transliteration - Arabic Code-Switched Translation --- # AraT5-base-title-generation # AraT5: Text-to-Text Transformers for Arabic Language Generation <img src="https://huggingface.co/UBC-NLP/AraT5-base/resolve/main/AraT5_CR_new.png" alt="AraT5" width="45%" height="35%" align="right"/> This is the repository accompanying our paper [AraT5: Text-to-Text Transformers for Arabic Language Understanding and Generation](https://aclanthology.org/2022.acl-long.47/). In this is the repository we Introduce AraT5<sub>MSA</sub>, AraT5<sub>Tweet</sub>, and AraT5: three powerful Arabic-specific text-to-text Transformer based models; --- # How to use AraT5 models Below is an example for fine-tuning AraT5-base for News Title Generation on the Aranews dataset ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("UBC-NLP/AraT5-base-title-generation") model = AutoModelForSeq2SeqLM.from_pretrained("UBC-NLP/AraT5-base-title-generation") Document = "تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة ." encoding = tokenizer.encode_plus(Document,pad_to_max_length=True, return_tensors="pt") input_ids, attention_masks = encoding["input_ids"], encoding["attention_mask"] outputs = model.generate( input_ids=input_ids, attention_mask=attention_masks, max_length=256, do_sample=True, top_k=120, top_p=0.95, early_stopping=True, num_return_sequences=5 ) for id, output in enumerate(outputs): title = tokenizer.decode(output, skip_special_tokens=True,clean_up_tokenization_spaces=True) print("title#"+str(id), title) ``` The input news document <div style="white-space : pre-wrap !important;word-break: break-word; direction:rtl; text-align: right"> تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة . <br> </div> The generated titles ``` title#0 غرفة الشرقية تختتم المرحلة الثانية من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة title#1 غرفة الشرقية تختتم الثاني من مبادرة تأهيل وتأهيل أبناء وبناتنا title#2 سعود بن نايف يختتم ثانى مبادراتها لتأهيل وتدريب أبناء وبنات المملكة title#3 أمير الشرقية يرعى اختتام برنامج برنامج تدريب أبناء وبنات المملكة title#4 سعود بن نايف يرعى اختتام مبادرة تأهيل وتدريب أبناء وبنات المملكة ``` # AraT5 Models Checkpoints AraT5 Pytorch and TensorFlow checkpoints are available on the Huggingface website for direct download and use ```exclusively for research```. ```For commercial use, please contact the authors via email @ (muhammad.mageed[at]ubc[dot]ca).``` \| Model \| Link \| \|---------\|:------------------:\| \| AraT5-base \| [https://huggingface.co/UBC-NLP/AraT5-base](https://huggingface.co/UBC-NLP/AraT5-base) \| \| AraT5-msa-base \| [https://huggingface.co/UBC-NLP/AraT5-msa-base](https://huggingface.co/UBC-NLP/AraT5-msa-base) \| \| AraT5-tweet-base \| [https://huggingface.co/UBC-NLP/AraT5-tweet-base](https://huggingface.co/UBC-NLP/AraT5-tweet-base) \| \| AraT5-msa-small \| [https://huggingface.co/UBC-NLP/AraT5-msa-small](https://huggingface.co/UBC-NLP/AraT5-msa-small) \| \| AraT5-tweet-small\| [https://huggingface.co/UBC-NLP/AraT5-tweet-small](https://huggingface.co/UBC-NLP/AraT5-tweet-small) \| # BibTex If you use our models (Arat5-base, Arat5-msa-base, Arat5-tweet-base, Arat5-msa-small, or Arat5-tweet-small ) for your scientific publication, or if you find the resources in this repository useful, please cite our paper as follows (to be updated): ```bibtex @inproceedings{nagoudi-etal-2022-arat5, title = "{A}ra{T}5: Text-to-Text Transformers for {A}rabic Language Generation", author = "Nagoudi, El Moatez Billah and Elmadany, AbdelRahim and Abdul-Mageed, Muhammad", booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = may, year = "2022", address = "Dublin, Ireland", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2022.acl-long.47", pages = "628--647", abstract = "Transfer learning with a unified Transformer framework (T5) that converts all language problems into a text-to-text format was recently proposed as a simple and effective transfer learning approach. Although a multilingual version of the T5 model (mT5) was also introduced, it is not clear how well it can fare on non-English tasks involving diverse data. To investigate this question, we apply mT5 on a language with a wide variety of dialects{--}Arabic. For evaluation, we introduce a novel benchmark for ARabic language GENeration (ARGEN), covering seven important tasks. For model comparison, we pre-train three powerful Arabic T5-style models and evaluate them on ARGEN. Although pre-trained with {\textasciitilde}49 less data, our new models perform significantly better than mT5 on all ARGEN tasks (in 52 out of 59 test sets) and set several new SOTAs. Our models also establish new SOTA on the recently-proposed, large Arabic language understanding evaluation benchmark ARLUE (Abdul-Mageed et al., 2021). Our new models are publicly available. We also link to ARGEN datasets through our repository: https://github.com/UBC-NLP/araT5.", } ``` ## Acknowledgments We gratefully acknowledge support from the Natural Sciences and Engineering Research Council of Canada, the Social Sciences and Humanities Research Council of Canada, Canadian Foundation for Innovation, [ComputeCanada](www.computecanada.ca) and [UBC ARC-Sockeye](https://doi.org/10.14288/SOCKEYE). We also thank the [Google TensorFlow Research Cloud (TFRC)](https://www.tensorflow.org/tfrc) program for providing us with free TPU access.	null	Non_BioNLP	# AraT5-base-title-generation # AraT5: Text-to-Text Transformers for Arabic Language Generation <img src="https://huggingface.co/UBC-NLP/AraT5-base/resolve/main/AraT5_CR_new.png" alt="AraT5" width="45%" height="35%" align="right"/> This is the repository accompanying our paper [AraT5: Text-to-Text Transformers for Arabic Language Understanding and Generation](https://aclanthology.org/2022.acl-long.47/). In this is the repository we Introduce AraT5<sub>MSA</sub>, AraT5<sub>Tweet</sub>, and AraT5: three powerful Arabic-specific text-to-text Transformer based models; --- # How to use AraT5 models Below is an example for fine-tuning AraT5-base for News Title Generation on the Aranews dataset ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("UBC-NLP/AraT5-base-title-generation") model = AutoModelForSeq2SeqLM.from_pretrained("UBC-NLP/AraT5-base-title-generation") Document = "تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة ." encoding = tokenizer.encode_plus(Document,pad_to_max_length=True, return_tensors="pt") input_ids, attention_masks = encoding["input_ids"], encoding["attention_mask"] outputs = model.generate( input_ids=input_ids, attention_mask=attention_masks, max_length=256, do_sample=True, top_k=120, top_p=0.95, early_stopping=True, num_return_sequences=5 ) for id, output in enumerate(outputs): title = tokenizer.decode(output, skip_special_tokens=True,clean_up_tokenization_spaces=True) print("title#"+str(id), title) ``` The input news document <div style="white-space : pre-wrap !important;word-break: break-word; direction:rtl; text-align: right"> تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة . <br> </div> The generated titles ``` title#0 غرفة الشرقية تختتم المرحلة الثانية من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة title#1 غرفة الشرقية تختتم الثاني من مبادرة تأهيل وتأهيل أبناء وبناتنا title#2 سعود بن نايف يختتم ثانى مبادراتها لتأهيل وتدريب أبناء وبنات المملكة title#3 أمير الشرقية يرعى اختتام برنامج برنامج تدريب أبناء وبنات المملكة title#4 سعود بن نايف يرعى اختتام مبادرة تأهيل وتدريب أبناء وبنات المملكة ``` # AraT5 Models Checkpoints AraT5 Pytorch and TensorFlow checkpoints are available on the Huggingface website for direct download and use ```exclusively for research```. ```For commercial use, please contact the authors via email @ (muhammad.mageed[at]ubc[dot]ca).``` \| Model \| Link \| \|---------\|:------------------:\| \| AraT5-base \| [https://huggingface.co/UBC-NLP/AraT5-base](https://huggingface.co/UBC-NLP/AraT5-base) \| \| AraT5-msa-base \| [https://huggingface.co/UBC-NLP/AraT5-msa-base](https://huggingface.co/UBC-NLP/AraT5-msa-base) \| \| AraT5-tweet-base \| [https://huggingface.co/UBC-NLP/AraT5-tweet-base](https://huggingface.co/UBC-NLP/AraT5-tweet-base) \| \| AraT5-msa-small \| [https://huggingface.co/UBC-NLP/AraT5-msa-small](https://huggingface.co/UBC-NLP/AraT5-msa-small) \| \| AraT5-tweet-small\| [https://huggingface.co/UBC-NLP/AraT5-tweet-small](https://huggingface.co/UBC-NLP/AraT5-tweet-small) \| # BibTex If you use our models (Arat5-base, Arat5-msa-base, Arat5-tweet-base, Arat5-msa-small, or Arat5-tweet-small ) for your scientific publication, or if you find the resources in this repository useful, please cite our paper as follows (to be updated): ```bibtex @inproceedings{nagoudi-etal-2022-arat5, title = "{A}ra{T}5: Text-to-Text Transformers for {A}rabic Language Generation", author = "Nagoudi, El Moatez Billah and Elmadany, AbdelRahim and Abdul-Mageed, Muhammad", booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = may, year = "2022", address = "Dublin, Ireland", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2022.acl-long.47", pages = "628--647", abstract = "Transfer learning with a unified Transformer framework (T5) that converts all language problems into a text-to-text format was recently proposed as a simple and effective transfer learning approach. Although a multilingual version of the T5 model (mT5) was also introduced, it is not clear how well it can fare on non-English tasks involving diverse data. To investigate this question, we apply mT5 on a language with a wide variety of dialects{--}Arabic. For evaluation, we introduce a novel benchmark for ARabic language GENeration (ARGEN), covering seven important tasks. For model comparison, we pre-train three powerful Arabic T5-style models and evaluate them on ARGEN. Although pre-trained with {\textasciitilde}49 less data, our new models perform significantly better than mT5 on all ARGEN tasks (in 52 out of 59 test sets) and set several new SOTAs. Our models also establish new SOTA on the recently-proposed, large Arabic language understanding evaluation benchmark ARLUE (Abdul-Mageed et al., 2021). Our new models are publicly available. We also link to ARGEN datasets through our repository: https://github.com/UBC-NLP/araT5.", } ``` ## Acknowledgments We gratefully acknowledge support from the Natural Sciences and Engineering Research Council of Canada, the Social Sciences and Humanities Research Council of Canada, Canadian Foundation for Innovation, [ComputeCanada](www.computecanada.ca) and [UBC ARC-Sockeye](https://doi.org/10.14288/SOCKEYE). We also thank the [Google TensorFlow Research Cloud (TFRC)](https://www.tensorflow.org/tfrc) program for providing us with free TPU access.	{"language": ["ar"], "tags": ["Arabic T5", "MSA", "Twitter", "Arabic Dialect", "Arabic Machine Translation", "Arabic Text Summarization", "Arabic News Title and Question Generation", "Arabic Paraphrasing and Transliteration", "Arabic Code-Switched Translation"]}	task	[ "TRANSLATION", "SUMMARIZATION", "PARAPHRASING" ]	44,624
Dhahlan2000/Simple_Translation-model-for-GPT-v13	Dhahlan2000	text2text-generation	[ "transformers", "tensorboard", "safetensors", "t5", "text2text-generation", "generated_from_trainer", "base_model:Dhahlan2000/Simple_Translation-model-for-GPT-v12", "base_model:finetune:Dhahlan2000/Simple_Translation-model-for-GPT-v12", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	2024-06-09T07:39:40Z	2024-06-09T07:40:02+00:00	4	0	--- base_model: Dhahlan2000/Simple_Translation-model-for-GPT-v12 metrics: - bleu tags: - generated_from_trainer model-index: - name: Simple_Translation-model-for-GPT-v13 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Simple_Translation-model-for-GPT-v13 This model is a fine-tuned version of [Dhahlan2000/Simple_Translation-model-for-GPT-v12](https://huggingface.co/Dhahlan2000/Simple_Translation-model-for-GPT-v12) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2668 - Bleu: 17.0937 - Gen Len: 18.5927 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Bleu \| Gen Len \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:-------:\|:-------:\| \| 0.3914 \| 1.0 \| 9282 \| 0.2714 \| 16.9453 \| 18.5893 \| \| 0.3774 \| 2.0 \| 18564 \| 0.2668 \| 17.0937 \| 18.5927 \| ### Framework versions - Transformers 4.41.2 - Pytorch 2.3.0+cu121 - Datasets 2.19.2 - Tokenizers 0.19.1	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Simple_Translation-model-for-GPT-v13 This model is a fine-tuned version of [Dhahlan2000/Simple_Translation-model-for-GPT-v12](https://huggingface.co/Dhahlan2000/Simple_Translation-model-for-GPT-v12) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2668 - Bleu: 17.0937 - Gen Len: 18.5927 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Bleu \| Gen Len \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:-------:\|:-------:\| \| 0.3914 \| 1.0 \| 9282 \| 0.2714 \| 16.9453 \| 18.5893 \| \| 0.3774 \| 2.0 \| 18564 \| 0.2668 \| 17.0937 \| 18.5927 \| ### Framework versions - Transformers 4.41.2 - Pytorch 2.3.0+cu121 - Datasets 2.19.2 - Tokenizers 0.19.1	{"base_model": "Dhahlan2000/Simple_Translation-model-for-GPT-v12", "metrics": ["bleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "Simple_Translation-model-for-GPT-v13", "results": []}]}	task	[ "TRANSLATION" ]	44,625
cross-encoder/stsb-roberta-base	cross-encoder	text-classification	[ "transformers", "pytorch", "jax", "safetensors", "roberta", "text-classification", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:05Z	2025-04-11T08:18:13+00:00	202,254	4	--- base_model: - FacebookAI/roberta-base datasets: - sentence-transformers/stsb language: - en library_name: sentence-transformers license: apache-2.0 pipeline_tag: text-ranking tags: - transformers --- # Cross-Encoder for Semantic Textual Similarity This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class. ## Training Data This model was trained on the [STS benchmark dataset](http://ixa2.si.ehu.eus/stswiki/index.php/STSbenchmark). The model will predict a score between 0 and 1 how for the semantic similarity of two sentences. ## Usage and Performance Pre-trained models can be used like this: ```python from sentence_transformers import CrossEncoder model = CrossEncoder('cross-encoder/stsb-roberta-base') scores = model.predict([('Sentence 1', 'Sentence 2'), ('Sentence 3', 'Sentence 4')]) ``` The model will predict scores for the pairs `('Sentence 1', 'Sentence 2')` and `('Sentence 3', 'Sentence 4')`. You can use this model also without sentence_transformers and by just using Transformers ``AutoModel`` class	null	Non_BioNLP	# Cross-Encoder for Semantic Textual Similarity This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class. ## Training Data This model was trained on the [STS benchmark dataset](http://ixa2.si.ehu.eus/stswiki/index.php/STSbenchmark). The model will predict a score between 0 and 1 how for the semantic similarity of two sentences. ## Usage and Performance Pre-trained models can be used like this: ```python from sentence_transformers import CrossEncoder model = CrossEncoder('cross-encoder/stsb-roberta-base') scores = model.predict([('Sentence 1', 'Sentence 2'), ('Sentence 3', 'Sentence 4')]) ``` The model will predict scores for the pairs `('Sentence 1', 'Sentence 2')` and `('Sentence 3', 'Sentence 4')`. You can use this model also without sentence_transformers and by just using Transformers ``AutoModel`` class	{"base_model": ["FacebookAI/roberta-base"], "datasets": ["sentence-transformers/stsb"], "language": ["en"], "library_name": "sentence-transformers", "license": "apache-2.0", "pipeline_tag": "text-ranking", "tags": ["transformers"]}	task	[ "SEMANTIC_SIMILARITY" ]	44,626
Helsinki-NLP/opus-mt-fi-guw	Helsinki-NLP	translation	[ "transformers", "pytorch", "tf", "marian", "text2text-generation", "translation", "fi", "guw", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2022-03-02T23:29:04Z	2023-08-16T11:34:37+00:00	48	0	--- license: apache-2.0 tags: - translation --- ### opus-mt-fi-guw * source languages: fi * target languages: guw * OPUS readme: [fi-guw](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fi-guw/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-08.zip](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.zip) * test set translations: [opus-2020-01-08.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.test.txt) * test set scores: [opus-2020-01-08.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.eval.txt) ## Benchmarks \| testset \| BLEU \| chr-F \| \|-----------------------\|-------\|-------\| \| JW300.fi.guw \| 32.4 \| 0.527 \|	null	Non_BioNLP	### opus-mt-fi-guw * source languages: fi * target languages: guw * OPUS readme: [fi-guw](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fi-guw/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-08.zip](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.zip) * test set translations: [opus-2020-01-08.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.test.txt) * test set scores: [opus-2020-01-08.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.eval.txt) ## Benchmarks \| testset \| BLEU \| chr-F \| \|-----------------------\|-------\|-------\| \| JW300.fi.guw \| 32.4 \| 0.527 \|	{"license": "apache-2.0", "tags": ["translation"]}	task	[ "TRANSLATION" ]	44,627
vishnupriyavr/flan-t5-movie-summary	vishnupriyavr	summarization	[ "peft", "art", "summarization", "en", "dataset:vishnupriyavr/wiki-movie-plots-with-summaries", "license:apache-2.0", "region:us" ]	2023-10-07T13:51:27Z	2023-10-07T20:34:48+00:00	2	0	--- datasets: - vishnupriyavr/wiki-movie-plots-with-summaries language: - en library_name: peft license: apache-2.0 metrics: - rouge pipeline_tag: summarization tags: - art ---	null	Non_BioNLP		{"datasets": ["vishnupriyavr/wiki-movie-plots-with-summaries"], "language": ["en"], "library_name": "peft", "license": "apache-2.0", "metrics": ["rouge"], "pipeline_tag": "summarization", "tags": ["art"]}	task	[ "SUMMARIZATION" ]	44,628
Luciferio/MiniLLM-finetuned	Luciferio	text-classification	[ "transformers", "pytorch", "safetensors", "bert", "text-classification", "generated_from_trainer", "dataset:emotion", "base_model:microsoft/MiniLM-L12-H384-uncased", "base_model:finetune:microsoft/MiniLM-L12-H384-uncased", "license:mit", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	2023-08-22T12:57:56Z	2023-08-22T14:39:58+00:00	14	0	--- base_model: microsoft/MiniLM-L12-H384-uncased datasets: - emotion license: mit metrics: - f1 tags: - generated_from_trainer model-index: - name: MiniLLM-finetuned results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion config: split split: validation args: split metrics: - type: f1 value: 0.922353805579638 name: F1 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # MiniLLM-finetuned This model is a fine-tuned version of [microsoft/MiniLM-L12-H384-uncased](https://huggingface.co/microsoft/MiniLM-L12-H384-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2932 - F1: 0.9224 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:------:\| \| No log \| 1.0 \| 2000 \| 0.4408 \| 0.8888 \| \| No log \| 2.0 \| 4000 \| 0.2932 \| 0.9224 \| ### Framework versions - Transformers 4.31.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # MiniLLM-finetuned This model is a fine-tuned version of [microsoft/MiniLM-L12-H384-uncased](https://huggingface.co/microsoft/MiniLM-L12-H384-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2932 - F1: 0.9224 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:------:\| \| No log \| 1.0 \| 2000 \| 0.4408 \| 0.8888 \| \| No log \| 2.0 \| 4000 \| 0.2932 \| 0.9224 \| ### Framework versions - Transformers 4.31.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3	{"base_model": "microsoft/MiniLM-L12-H384-uncased", "datasets": ["emotion"], "license": "mit", "metrics": ["f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "MiniLLM-finetuned", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "config": "split", "split": "validation", "args": "split"}, "metrics": [{"type": "f1", "value": 0.922353805579638, "name": "F1"}]}]}]}	task	[ "TEXT_CLASSIFICATION" ]	44,629
ccdv/lsg-bart-base-4096-wcep	ccdv	summarization	[ "transformers", "pytorch", "bart", "text2text-generation", "summarization", "custom_code", "en", "dataset:ccdv/WCEP-10", "arxiv:2210.15497", "autotrain_compatible", "region:us" ]	2022-05-25T11:09:11Z	2023-12-17T21:10:24+00:00	50	2	--- datasets: - ccdv/WCEP-10 language: - en metrics: - rouge tags: - summarization model-index: - name: ccdv/lsg-bart-base-4096-wcep results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> Transformers >= 4.36.1\ This model relies on a custom modeling file, you need to add trust_remote_code=True\ See [\#13467](https://github.com/huggingface/transformers/pull/13467) LSG ArXiv [paper](https://arxiv.org/abs/2210.15497). \ Github/conversion script is available at this [link](https://github.com/ccdv-ai/convert_checkpoint_to_lsg). ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline tokenizer = AutoTokenizer.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) model = AutoModelForSeq2SeqLM.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) text = "Replace by what you want." pipe = pipeline("text2text-generation", model=model, tokenizer=tokenizer, device=0) generated_text = pipe(text, truncation=True, max_length=64, no_repeat_ngram_size=7) ``` # ccdv/lsg-bart-base-4096-wcep This model is a fine-tuned version of [ccdv/lsg-bart-base-4096](https://huggingface.co/ccdv/lsg-bart-base-4096) on the [ccdv/WCEP-10 roberta](https://huggingface.co/datasets/ccdv/WCEP-10) dataset. \ It achieves the following results on the test set: \| Length \| Sparse Type \| Block Size \| Sparsity \| Connexions \| R1 \| R2 \| RL \| RLsum \| \|:------ \|:------------ \|:---------- \|:-------- \| :--------- \|:----- \|:----- \|:----- \|:----- \| \| 4096 \| Local \| 256 \| 0 \| 768 \| 46.02 \| 24.23 \| 37.38 \| 38.72 \| \| 4096 \| Local \| 128 \| 0 \| 384 \| 45.43 \| 23.86 \| 36.94 \| 38.30 \| \| 4096 \| Pooling \| 128 \| 4 \| 644 \| 45.36 \| 23.61 \| 36.75 \| 38.06 \| \| 4096 \| Stride \| 128 \| 4 \| 644 \| 45.87 \| 24.31 \| 37.41 \| 38.70 \| \| 4096 \| Block Stride \| 128 \| 4 \| 644 \| 45.78 \| 24.16 \| 37.20 \| 38.48 \| \| 4096 \| Norm \| 128 \| 4 \| 644 \| 45.34 \| 23.39 \| 36.47 \| 37.78 \| \| 4096 \| LSH \| 128 \| 4 \| 644 \| 45.15 \| 23.53 \| 36.74 \| 38.02 \| With smaller block size (lower ressources): \| Length \| Sparse Type \| Block Size \| Sparsity \| Connexions \| R1 \| R2 \| RL \| RLsum \| \|:------ \|:------------ \|:---------- \|:-------- \| :--------- \|:----- \|:----- \|:----- \|:----- \| \| 4096 \| Local \| 64 \| 0 \| 192 \| 44.48 \| 22.98 \| 36.20 \| 37.52 \| \| 4096 \| Local \| 32 \| 0 \| 96 \| 43.60 \| 22.17 \| 35.61 \| 36.66 \| \| 4096 \| Pooling \| 32 \| 4 \| 160 \| 43.91 \| 22.41 \| 35.80 \| 36.92 \| \| 4096 \| Stride \| 32 \| 4 \| 160 \| 44.62 \| 23.11 \| 36.32 \| 37.53 \| \| 4096 \| Block Stride \| 32 \| 4 \| 160 \| 44.47 \| 23.02 \| 36.28 \| 37.46 \| \| 4096 \| Norm \| 32 \| 4 \| 160 \| 44.45 \| 23.03 \| 36.10 \| 37.33 \| \| 4096 \| LSH \| 32 \| 4 \| 160 \| 43.87 \| 22.50 \| 35.75 \| 36.93 \| ## Model description The model relies on Local-Sparse-Global attention to handle long sequences: ![attn](attn.png) The model has about ~145 millions parameters (6 encoder layers - 6 decoder layers). \ The model is warm started from BART-base, converted to handle long sequences (encoder only) and fine tuned. ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 8e-05 - train_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 10.0 ### Generate hyperparameters The following hyperparameters were used during generation: - dataset_name: ccdv/WCEP-10 - dataset_config_name: roberta - eval_batch_size: 8 - eval_samples: 1022 - early_stopping: True - ignore_pad_token_for_loss: True - length_penalty: 2.0 - max_length: 64 - min_length: 0 - num_beams: 5 - no_repeat_ngram_size: None - seed: 123 ### Framework versions - Transformers 4.18.0 - Pytorch 1.10.1+cu102 - Datasets 2.1.0 - Tokenizers 0.11.6	null	Non_BioNLP	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> Transformers >= 4.36.1\ This model relies on a custom modeling file, you need to add trust_remote_code=True\ See [\#13467](https://github.com/huggingface/transformers/pull/13467) LSG ArXiv [paper](https://arxiv.org/abs/2210.15497). \ Github/conversion script is available at this [link](https://github.com/ccdv-ai/convert_checkpoint_to_lsg). ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline tokenizer = AutoTokenizer.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) model = AutoModelForSeq2SeqLM.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) text = "Replace by what you want." pipe = pipeline("text2text-generation", model=model, tokenizer=tokenizer, device=0) generated_text = pipe(text, truncation=True, max_length=64, no_repeat_ngram_size=7) ``` # ccdv/lsg-bart-base-4096-wcep This model is a fine-tuned version of [ccdv/lsg-bart-base-4096](https://huggingface.co/ccdv/lsg-bart-base-4096) on the [ccdv/WCEP-10 roberta](https://huggingface.co/datasets/ccdv/WCEP-10) dataset. \ It achieves the following results on the test set: \| Length \| Sparse Type \| Block Size \| Sparsity \| Connexions \| R1 \| R2 \| RL \| RLsum \| \|:------ \|:------------ \|:---------- \|:-------- \| :--------- \|:----- \|:----- \|:----- \|:----- \| \| 4096 \| Local \| 256 \| 0 \| 768 \| 46.02 \| 24.23 \| 37.38 \| 38.72 \| \| 4096 \| Local \| 128 \| 0 \| 384 \| 45.43 \| 23.86 \| 36.94 \| 38.30 \| \| 4096 \| Pooling \| 128 \| 4 \| 644 \| 45.36 \| 23.61 \| 36.75 \| 38.06 \| \| 4096 \| Stride \| 128 \| 4 \| 644 \| 45.87 \| 24.31 \| 37.41 \| 38.70 \| \| 4096 \| Block Stride \| 128 \| 4 \| 644 \| 45.78 \| 24.16 \| 37.20 \| 38.48 \| \| 4096 \| Norm \| 128 \| 4 \| 644 \| 45.34 \| 23.39 \| 36.47 \| 37.78 \| \| 4096 \| LSH \| 128 \| 4 \| 644 \| 45.15 \| 23.53 \| 36.74 \| 38.02 \| With smaller block size (lower ressources): \| Length \| Sparse Type \| Block Size \| Sparsity \| Connexions \| R1 \| R2 \| RL \| RLsum \| \|:------ \|:------------ \|:---------- \|:-------- \| :--------- \|:----- \|:----- \|:----- \|:----- \| \| 4096 \| Local \| 64 \| 0 \| 192 \| 44.48 \| 22.98 \| 36.20 \| 37.52 \| \| 4096 \| Local \| 32 \| 0 \| 96 \| 43.60 \| 22.17 \| 35.61 \| 36.66 \| \| 4096 \| Pooling \| 32 \| 4 \| 160 \| 43.91 \| 22.41 \| 35.80 \| 36.92 \| \| 4096 \| Stride \| 32 \| 4 \| 160 \| 44.62 \| 23.11 \| 36.32 \| 37.53 \| \| 4096 \| Block Stride \| 32 \| 4 \| 160 \| 44.47 \| 23.02 \| 36.28 \| 37.46 \| \| 4096 \| Norm \| 32 \| 4 \| 160 \| 44.45 \| 23.03 \| 36.10 \| 37.33 \| \| 4096 \| LSH \| 32 \| 4 \| 160 \| 43.87 \| 22.50 \| 35.75 \| 36.93 \| ## Model description The model relies on Local-Sparse-Global attention to handle long sequences: ![attn](attn.png) The model has about ~145 millions parameters (6 encoder layers - 6 decoder layers). \ The model is warm started from BART-base, converted to handle long sequences (encoder only) and fine tuned. ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 8e-05 - train_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 10.0 ### Generate hyperparameters The following hyperparameters were used during generation: - dataset_name: ccdv/WCEP-10 - dataset_config_name: roberta - eval_batch_size: 8 - eval_samples: 1022 - early_stopping: True - ignore_pad_token_for_loss: True - length_penalty: 2.0 - max_length: 64 - min_length: 0 - num_beams: 5 - no_repeat_ngram_size: None - seed: 123 ### Framework versions - Transformers 4.18.0 - Pytorch 1.10.1+cu102 - Datasets 2.1.0 - Tokenizers 0.11.6	{"datasets": ["ccdv/WCEP-10"], "language": ["en"], "metrics": ["rouge"], "tags": ["summarization"], "model-index": [{"name": "ccdv/lsg-bart-base-4096-wcep", "results": []}]}	task	[ "SUMMARIZATION" ]	44,630
Shankhdhar/ecommerce_query_classifier	Shankhdhar	text-classification	[ "sentence-transformers", "safetensors", "bert", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	2023-12-03T18:13:32Z	2023-12-03T18:25:32+00:00	50	0	--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- #Shankhdhar/ecommerce_query_classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Shankhdhar/ecommerce_query_classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	null	Non_BioNLP	#Shankhdhar/ecommerce_query_classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Shankhdhar/ecommerce_query_classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```	{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}	task	[ "TEXT_CLASSIFICATION" ]	44,631

IAmSkyDra/BART_Translation_Finetune_v0

IAmSkyDra

text2text-generation

[ "transformers", "safetensors", "mbart", "text2text-generation", "generated_from_trainer", "base_model:vinai/bartpho-syllable", "base_model:finetune:vinai/bartpho-syllable", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2025-01-29T07:54:22Z

2025-01-29T09:14:59+00:00

17

0

--- base_model: vinai/bartpho-syllable library_name: transformers license: mit metrics: - sacrebleu tags: - generated_from_trainer model-index: - name: BART_Translation_Finetune_v0 results: [] ---  # BART_Translation_Finetune_v0 This model is a fine-tuned version of [vinai/bartpho-syllable](https://huggingface.co/vinai/bartpho-syllable) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.4390 - Sacrebleu: 6.4288 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 100 - eval_batch_size: 100 - seed: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 15 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Sacrebleu | |:-------------:|:-----:|:----:|:---------------:|:---------:| | 0.7274 | 1.0 | 468 | 0.6396 | 1.4512 | | 0.628 | 2.0 | 936 | 0.5901 | 2.1680 | | 0.5744 | 3.0 | 1404 | 0.5431 | 3.0785 | | 0.5348 | 4.0 | 1872 | 0.5141 | 3.9172 | | 0.5044 | 5.0 | 2340 | 0.4905 | 4.3428 | | 0.4773 | 6.0 | 2808 | 0.4758 | 4.8562 | | 0.4575 | 7.0 | 3276 | 0.4647 | 5.2799 | | 0.4372 | 8.0 | 3744 | 0.4572 | 5.5700 | | 0.429 | 9.0 | 4212 | 0.4499 | 5.8456 | | 0.4118 | 10.0 | 4680 | 0.4455 | 6.0842 | | 0.4056 | 11.0 | 5148 | 0.4424 | 6.2403 | | 0.3924 | 12.0 | 5616 | 0.4403 | 6.2796 | | 0.3858 | 13.0 | 6084 | 0.4396 | 6.3191 | | 0.386 | 14.0 | 6552 | 0.4389 | 6.4120 | | 0.3809 | 15.0 | 7020 | 0.4390 | 6.4288 | ### Framework versions - Transformers 4.48.1 - Pytorch 2.5.1+cu124 - Datasets 3.2.0 - Tokenizers 0.21.0

null

Non_BioNLP

# BART_Translation_Finetune_v0 This model is a fine-tuned version of [vinai/bartpho-syllable](https://huggingface.co/vinai/bartpho-syllable) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.4390 - Sacrebleu: 6.4288 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 100 - eval_batch_size: 100 - seed: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 15 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Sacrebleu | |:-------------:|:-----:|:----:|:---------------:|:---------:| | 0.7274 | 1.0 | 468 | 0.6396 | 1.4512 | | 0.628 | 2.0 | 936 | 0.5901 | 2.1680 | | 0.5744 | 3.0 | 1404 | 0.5431 | 3.0785 | | 0.5348 | 4.0 | 1872 | 0.5141 | 3.9172 | | 0.5044 | 5.0 | 2340 | 0.4905 | 4.3428 | | 0.4773 | 6.0 | 2808 | 0.4758 | 4.8562 | | 0.4575 | 7.0 | 3276 | 0.4647 | 5.2799 | | 0.4372 | 8.0 | 3744 | 0.4572 | 5.5700 | | 0.429 | 9.0 | 4212 | 0.4499 | 5.8456 | | 0.4118 | 10.0 | 4680 | 0.4455 | 6.0842 | | 0.4056 | 11.0 | 5148 | 0.4424 | 6.2403 | | 0.3924 | 12.0 | 5616 | 0.4403 | 6.2796 | | 0.3858 | 13.0 | 6084 | 0.4396 | 6.3191 | | 0.386 | 14.0 | 6552 | 0.4389 | 6.4120 | | 0.3809 | 15.0 | 7020 | 0.4390 | 6.4288 | ### Framework versions - Transformers 4.48.1 - Pytorch 2.5.1+cu124 - Datasets 3.2.0 - Tokenizers 0.21.0

{"base_model": "vinai/bartpho-syllable", "library_name": "transformers", "license": "mit", "metrics": ["sacrebleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "BART_Translation_Finetune_v0", "results": []}]}

task

[ "TRANSLATION" ]

44,529

zedfum/arman-longformer-8k-finetuned-ensani

zedfum

summarization

[ "transformers", "pytorch", "safetensors", "pegasus", "text2text-generation", "summarization", "fa", "dataset:zedfum/long-summarization-persian", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-06-19T07:32:56Z

2023-07-18T16:31:10+00:00

68

2

--- datasets: - zedfum/long-summarization-persian language: - fa pipeline_tag: summarization ---  Persian Long Text Summarization This Model Accept 8K Tokens ## ⚡️ Quickstart ``` from transformers import AutoTokenizer from transformers import pipeline summarizer = pipeline("summarization", model="zedfum/arman-longformer-8k-finetuned-ensani", tokenizer="zedfum/arman-longformer-8k-finetuned-ensani" , device=0) text_to_summarize="" summarizer(text_to_summarize, min_length=5, max_length=512, truncation=True) ```

null

Non_BioNLP

Persian Long Text Summarization This Model Accept 8K Tokens ## ⚡️ Quickstart ``` from transformers import AutoTokenizer from transformers import pipeline summarizer = pipeline("summarization", model="zedfum/arman-longformer-8k-finetuned-ensani", tokenizer="zedfum/arman-longformer-8k-finetuned-ensani" , device=0) text_to_summarize="" summarizer(text_to_summarize, min_length=5, max_length=512, truncation=True) ```

{"datasets": ["zedfum/long-summarization-persian"], "language": ["fa"], "pipeline_tag": "summarization"}

task

[ "SUMMARIZATION" ]

44,530

datlaaaaaaa/81d3479a-ff61-4e88-8c7b-3c3d80ee47bf

datlaaaaaaa

null

[ "peft", "safetensors", "llama", "axolotl", "generated_from_trainer", "base_model:unsloth/llama-3-8b", "base_model:adapter:unsloth/llama-3-8b", "license:llama3", "8-bit", "bitsandbytes", "region:us" ]

2025-01-18T12:15:07Z

2025-01-18T12:31:41+00:00

1

0

--- base_model: unsloth/llama-3-8b library_name: peft license: llama3 tags: - axolotl - generated_from_trainer model-index: - name: 81d3479a-ff61-4e88-8c7b-3c3d80ee47bf results: [] ---  [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: unsloth/llama-3-8b bf16: auto chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 1b99c07a98e6036a_train_data.json ds_type: json format: custom path: /workspace/input_data/1b99c07a98e6036a_train_data.json type: field_input: choices field_instruction: input_translation field_output: input format: '{instruction} {input}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null deepspeed: null early_stopping_patience: null eval_max_new_tokens: 128 eval_table_size: null evals_per_epoch: 1 flash_attention: true fp16: null fsdp: null fsdp_config: null gradient_accumulation_steps: 4 gradient_checkpointing: true gradient_clipping: 1.0 group_by_length: false hub_model_id: datlaaaaaaa/81d3479a-ff61-4e88-8c7b-3c3d80ee47bf hub_repo: null hub_strategy: end hub_token: null learning_rate: 5.0e-05 load_in_4bit: true load_in_8bit: true local_rank: null logging_steps: 1 lora_alpha: 16 lora_dropout: 0.05 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 8 lora_target_linear: true lr_scheduler: cosine max_steps: 200 micro_batch_size: 2 mlflow_experiment_name: /tmp/1b99c07a98e6036a_train_data.json model_type: AutoModelForCausalLM num_epochs: 1 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false saves_per_epoch: 1 sequence_len: 1024 strict: false tf32: false tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.05 wandb_entity: null wandb_mode: online wandb_name: 582a5157-bb3b-4681-835d-d17ab3d65f4d wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 582a5157-bb3b-4681-835d-d17ab3d65f4d warmup_steps: 5 weight_decay: 0.01 xformers_attention: true ``` </details> # 81d3479a-ff61-4e88-8c7b-3c3d80ee47bf This model is a fine-tuned version of [unsloth/llama-3-8b](https://huggingface.co/unsloth/llama-3-8b) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.2841 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 8 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 5 - training_steps: 200 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | 0.2444 | 0.7048 | 200 | 0.2841 | ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1

null

Non_BioNLP

[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: unsloth/llama-3-8b bf16: auto chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 1b99c07a98e6036a_train_data.json ds_type: json format: custom path: /workspace/input_data/1b99c07a98e6036a_train_data.json type: field_input: choices field_instruction: input_translation field_output: input format: '{instruction} {input}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null deepspeed: null early_stopping_patience: null eval_max_new_tokens: 128 eval_table_size: null evals_per_epoch: 1 flash_attention: true fp16: null fsdp: null fsdp_config: null gradient_accumulation_steps: 4 gradient_checkpointing: true gradient_clipping: 1.0 group_by_length: false hub_model_id: datlaaaaaaa/81d3479a-ff61-4e88-8c7b-3c3d80ee47bf hub_repo: null hub_strategy: end hub_token: null learning_rate: 5.0e-05 load_in_4bit: true load_in_8bit: true local_rank: null logging_steps: 1 lora_alpha: 16 lora_dropout: 0.05 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 8 lora_target_linear: true lr_scheduler: cosine max_steps: 200 micro_batch_size: 2 mlflow_experiment_name: /tmp/1b99c07a98e6036a_train_data.json model_type: AutoModelForCausalLM num_epochs: 1 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false saves_per_epoch: 1 sequence_len: 1024 strict: false tf32: false tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.05 wandb_entity: null wandb_mode: online wandb_name: 582a5157-bb3b-4681-835d-d17ab3d65f4d wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 582a5157-bb3b-4681-835d-d17ab3d65f4d warmup_steps: 5 weight_decay: 0.01 xformers_attention: true ``` </details> # 81d3479a-ff61-4e88-8c7b-3c3d80ee47bf This model is a fine-tuned version of [unsloth/llama-3-8b](https://huggingface.co/unsloth/llama-3-8b) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.2841 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 8 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 5 - training_steps: 200 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | 0.2444 | 0.7048 | 200 | 0.2841 | ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1

{"base_model": "unsloth/llama-3-8b", "library_name": "peft", "license": "llama3", "tags": ["axolotl", "generated_from_trainer"], "model-index": [{"name": "81d3479a-ff61-4e88-8c7b-3c3d80ee47bf", "results": []}]}

task

[ "TRANSLATION" ]

44,531

navjordj/flan-t5-large_en-no

navjordj

text2text-generation

[ "transformers", "pytorch", "t5", "text2text-generation", "generated_from_trainer", "en", "no", "dataset:bible_para", "license:apache-2.0", "model-index", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2022-12-02T19:55:21Z

2022-12-03T15:33:10+00:00

16

0

--- datasets: - bible_para language: - en - 'no' license: apache-2.0 metrics: - bleu tags: - generated_from_trainer model-index: - name: flan-t5-large_en-no results: - task: type: translation name: Translation dataset: name: bible_para en-no type: bible_para config: en-no split: train args: en-no metrics: - type: bleu value: 36.7184 name: Bleu ---  # flan-t5-large_en-no This model is a fine-tuned version of [google/flan-t5-large](https://huggingface.co/google/flan-t5-large) on the bible_para en-no dataset. It achieves the following results on the evaluation set: - Loss: 0.6335 - Bleu: 36.7184 - Gen Len: 64.6249 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5.0 ### Training results ### Framework versions - Transformers 4.26.0.dev0 - Pytorch 1.13.0 - Datasets 2.7.1 - Tokenizers 0.13.2

null

Non_BioNLP

# flan-t5-large_en-no This model is a fine-tuned version of [google/flan-t5-large](https://huggingface.co/google/flan-t5-large) on the bible_para en-no dataset. It achieves the following results on the evaluation set: - Loss: 0.6335 - Bleu: 36.7184 - Gen Len: 64.6249 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5.0 ### Training results ### Framework versions - Transformers 4.26.0.dev0 - Pytorch 1.13.0 - Datasets 2.7.1 - Tokenizers 0.13.2

{"datasets": ["bible_para"], "language": ["en", "no"], "license": "apache-2.0", "metrics": ["bleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "flan-t5-large_en-no", "results": [{"task": {"type": "translation", "name": "Translation"}, "dataset": {"name": "bible_para en-no", "type": "bible_para", "config": "en-no", "split": "train", "args": "en-no"}, "metrics": [{"type": "bleu", "value": 36.7184, "name": "Bleu"}]}]}]}

task

[ "TRANSLATION" ]

44,532

dabraldeepti25/legal-ft-v0-updated

dabraldeepti25

sentence-similarity

[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:156", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss", "arxiv:1908.10084", "arxiv:2205.13147", "arxiv:1705.00652", "base_model:Snowflake/snowflake-arctic-embed-l", "base_model:finetune:Snowflake/snowflake-arctic-embed-l", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

2025-02-18T22:13:53Z

2025-02-18T22:14:27+00:00

13

0

--- base_model: Snowflake/snowflake-arctic-embed-l library_name: sentence-transformers metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:156 - loss:MatryoshkaLoss - loss:MultipleNegativesRankingLoss widget: - source_sentence: What significant change occurred in the AI landscape regarding models surpassing GPT-4 in the past twelve months? sentences: - 'Except... you can run generated code to see if it’s correct. And with patterns like ChatGPT Code Interpreter the LLM can execute the code itself, process the error message, then rewrite it and keep trying until it works! So hallucination is a much lesser problem for code generation than for anything else. If only we had the equivalent of Code Interpreter for fact-checking natural language! How should we feel about this as software engineers? On the one hand, this feels like a threat: who needs a programmer if ChatGPT can write code for you?' - 'The GPT-4 barrier was comprehensively broken In my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t? I’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.' - 'If you think about what they do, this isn’t such a big surprise. The grammar rules of programming languages like Python and JavaScript are massively less complicated than the grammar of Chinese, Spanish or English. It’s still astonishing to me how effective they are though. One of the great weaknesses of LLMs is their tendency to hallucinate—to imagine things that don’t correspond to reality. You would expect this to be a particularly bad problem for code—if an LLM hallucinates a method that doesn’t exist, the code should be useless.' - source_sentence: How does Claude enable users to interact with applications created through its interface? sentences: - 'OpenAI made GPT-4o free for all users in May, and Claude 3.5 Sonnet was freely available from its launch in June. This was a momentus change, because for the previous year free users had mostly been restricted to GPT-3.5 level models, meaning new users got a very inaccurate mental model of what a capable LLM could actually do. That era appears to have ended, likely permanently, with OpenAI’s launch of ChatGPT Pro. This $200/month subscription service is the only way to access their most capable model, o1 Pro. Since the trick behind the o1 series (and the future models it will undoubtedly inspire) is to expend more compute time to get better results, I don’t think those days of free access to the best available models are likely to return.' - 'We already knew LLMs were spookily good at writing code. If you prompt them right, it turns out they can build you a full interactive application using HTML, CSS and JavaScript (and tools like React if you wire up some extra supporting build mechanisms)—often in a single prompt. Anthropic kicked this idea into high gear when they released Claude Artifacts, a groundbreaking new feature that was initially slightly lost in the noise due to being described half way through their announcement of the incredible Claude 3.5 Sonnet. With Artifacts, Claude can write you an on-demand interactive application and then let you use it directly inside the Claude interface. Here’s my Extract URLs app, entirely generated by Claude:' - "Industry’s Tardy Response to the AI Prompt Injection Vulnerability on RedMonk\ \ Conversations\n\n\nPosted 31st December 2023 at 11:59 pm · Follow me on Mastodon\ \ or Twitter or subscribe to my newsletter\n\n\nMore recent articles\n\nLLM 0.22,\ \ the annotated release notes - 17th February 2025\nRun LLMs on macOS using llm-mlx\ \ and Apple's MLX framework - 15th February 2025\nURL-addressable Pyodide Python\ \ environments - 13th February 2025\n\n\n \n\n\nThis is Stuff we figured out about\ \ AI in 2023 by Simon Willison, posted on 31st December 2023.\n\nPart of series\ \ LLMs annual review\n\nStuff we figured out about AI in 2023 - Dec. 31, 2023,\ \ 11:59 p.m. \nThings we learned about LLMs in 2024 - Dec. 31, 2024, 6:07 p.m.\ \ \n\n\n\n blogging\n 69" - source_sentence: What incident involving Google Search is mentioned in the context, and what was the nature of the misinformation? sentences: - 'My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment). That same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:' - 'Terminology aside, I remain skeptical as to their utility based, once again, on the challenge of gullibility. LLMs believe anything you tell them. Any systems that attempts to make meaningful decisions on your behalf will run into the same roadblock: how good is a travel agent, or a digital assistant, or even a research tool if it can’t distinguish truth from fiction? Just the other day Google Search was caught serving up an entirely fake description of the non-existant movie “Encanto 2”. It turned out to be summarizing an imagined movie listing from a fan fiction wiki.' - 'On the other hand, as software engineers we are better placed to take advantage of this than anyone else. We’ve all been given weird coding interns—we can use our deep knowledge to prompt them to solve coding problems more effectively than anyone else can. The ethics of this space remain diabolically complex In September last year Andy Baio and I produced the first major story on the unlicensed training data behind Stable Diffusion. Since then, almost every major LLM (and most of the image generation models) have also been trained on unlicensed data.' - source_sentence: What are the limitations of Apple's LLM features compared to frontier LLMs, according to the context? sentences: - 'DeepSeek v3 is a huge 685B parameter model—one of the largest openly licensed models currently available, significantly bigger than the largest of Meta’s Llama series, Llama 3.1 405B. Benchmarks put it up there with Claude 3.5 Sonnet. Vibe benchmarks (aka the Chatbot Arena) currently rank it 7th, just behind the Gemini 2.0 and OpenAI 4o/o1 models. This is by far the highest ranking openly licensed model. The really impressive thing about DeepSeek v3 is the training cost. The model was trained on 2,788,000 H800 GPU hours at an estimated cost of $5,576,000. Llama 3.1 405B trained 30,840,000 GPU hours—11x that used by DeepSeek v3, for a model that benchmarks slightly worse.' - 'An interesting point of comparison here could be the way railways rolled out around the world in the 1800s. Constructing these required enormous investments and had a massive environmental impact, and many of the lines that were built turned out to be unnecessary—sometimes multiple lines from different companies serving the exact same routes! The resulting bubbles contributed to several financial crashes, see Wikipedia for Panic of 1873, Panic of 1893, Panic of 1901 and the UK’s Railway Mania. They left us with a lot of useful infrastructure and a great deal of bankruptcies and environmental damage. The year of slop' - 'Now that those features are rolling out they’re pretty weak. As an LLM power-user I know what these models are capable of, and Apple’s LLM features offer a pale imitation of what a frontier LLM can do. Instead we’re getting notification summaries that misrepresent news headlines and writing assistant tools that I’ve not found useful at all. Genmoji are kind of fun though. The rise of inference-scaling “reasoning” models The most interesting development in the final quarter of 2024 was the introduction of a new shape of LLM, exemplified by OpenAI’s o1 models—initially released as o1-preview and o1-mini on September 12th.' - source_sentence: What new feature was introduced in ChatGPT's voice mode in December? sentences: - The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did. - 'The two main categories I see are people who think AI agents are obviously things that go and act on your behalf—the travel agent model—and people who think in terms of LLMs that have been given access to tools which they can run in a loop as part of solving a problem. The term “autonomy” is often thrown into the mix too, again without including a clear definition. (I also collected 211 definitions on Twitter a few months ago—here they are in Datasette Lite—and had gemini-exp-1206 attempt to summarize them.) Whatever the term may mean, agents still have that feeling of perpetually “coming soon”.' - 'The GPT-4 barrier was comprehensively broken In my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t? I’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.' model-index: - name: SentenceTransformer based on Snowflake/snowflake-arctic-embed-l results: - task: type: information-retrieval name: Information Retrieval dataset: name: Unknown type: unknown metrics: - type: cosine_accuracy@1 value: 0.9166666666666666 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 1.0 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 1.0 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 1.0 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.9166666666666666 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.3333333333333333 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.20000000000000004 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.10000000000000002 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.9166666666666666 name: Cosine Recall@1 - type: cosine_recall@3 value: 1.0 name: Cosine Recall@3 - type: cosine_recall@5 value: 1.0 name: Cosine Recall@5 - type: cosine_recall@10 value: 1.0 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.9692441461309548 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.9583333333333334 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.9583333333333334 name: Cosine Map@100 --- # SentenceTransformer based on Snowflake/snowflake-arctic-embed-l This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 1024 dimensions - **Similarity Function:** Cosine Similarity    ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("dabraldeepti25/legal-ft-v0-updated") # Run inference sentences = [ "What new feature was introduced in ChatGPT's voice mode in December?", 'The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did.', 'The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 1024] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Information Retrieval * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.9167 | | cosine_accuracy@3 | 1.0 | | cosine_accuracy@5 | 1.0 | | cosine_accuracy@10 | 1.0 | | cosine_precision@1 | 0.9167 | | cosine_precision@3 | 0.3333 | | cosine_precision@5 | 0.2 | | cosine_precision@10 | 0.1 | | cosine_recall@1 | 0.9167 | | cosine_recall@3 | 1.0 | | cosine_recall@5 | 1.0 | | cosine_recall@10 | 1.0 | | **cosine_ndcg@10** | **0.9692** | | cosine_mrr@10 | 0.9583 | | cosine_map@100 | 0.9583 |   ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 156 training samples * Columns: <code>sentence_0</code> and <code>sentence_1</code> * Approximate statistics based on the first 156 samples: | | sentence_0 | sentence_1 | |:--------|:-----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| | type | string | string | | details | <ul><li>min: 13 tokens</li><li>mean: 20.17 tokens</li><li>max: 34 tokens</li></ul> | <ul><li>min: 43 tokens</li><li>mean: 135.18 tokens</li><li>max: 214 tokens</li></ul> | * Samples: | sentence_0 | sentence_1 | |:-------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | <code>What is the significance of prompt engineering in DALL-E 3?</code> | <code>Now add a walrus: Prompt engineering in DALL-E 3 32.8k 41.2k Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive 32.5k 38.2k ChatGPT can’t access the internet, even though it really looks like it can 30.5k 34.2k Stanford Alpaca, and the acceleration of on-device large language model development 29.7k 35.7k Run Llama 2 on your own Mac using LLM and Homebrew 27.9k 33.6k Midjourney 5.1 26.7k 33.4k Think of language models like ChatGPT as a “calculator for words” 25k 31.8k Multi-modal prompt injection image attacks against GPT-4V 23.7k 27.4k</code> | | <code>How does the vicuna-7b Large Language Model operate within a browser?</code> | <code>Now add a walrus: Prompt engineering in DALL-E 3 32.8k 41.2k Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive 32.5k 38.2k ChatGPT can’t access the internet, even though it really looks like it can 30.5k 34.2k Stanford Alpaca, and the acceleration of on-device large language model development 29.7k 35.7k Run Llama 2 on your own Mac using LLM and Homebrew 27.9k 33.6k Midjourney 5.1 26.7k 33.4k Think of language models like ChatGPT as a “calculator for words” 25k 31.8k Multi-modal prompt injection image attacks against GPT-4V 23.7k 27.4k</code> | | <code>What model of MacBook Pro is being used in the context, and what is its storage capacity?</code> | <code>My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment). That same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:</code> | * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 768, 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `num_train_epochs`: 10 - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `torch_empty_cache_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.0 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: False - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: None - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `include_for_metrics`: [] - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin </details> ### Training Logs | Epoch | Step | cosine_ndcg@10 | |:-----:|:----:|:--------------:| | 1.0 | 16 | 0.9692 | | 2.0 | 32 | 0.9692 | | 3.0 | 48 | 1.0 | | 3.125 | 50 | 1.0 | | 4.0 | 64 | 1.0 | | 5.0 | 80 | 0.9692 | | 6.0 | 96 | 0.9692 | | 6.25 | 100 | 0.9692 | | 7.0 | 112 | 0.9692 | | 8.0 | 128 | 0.9692 | | 9.0 | 144 | 0.9692 | | 9.375 | 150 | 0.9692 | | 10.0 | 160 | 0.9692 | ### Framework Versions - Python: 3.11.11 - Sentence Transformers: 3.4.1 - Transformers: 4.48.3 - PyTorch: 2.5.1+cu124 - Accelerate: 1.3.0 - Datasets: 3.3.1 - Tokenizers: 0.21.0 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

null

Non_BioNLP

# SentenceTransformer based on Snowflake/snowflake-arctic-embed-l This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 1024 dimensions - **Similarity Function:** Cosine Similarity    ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("dabraldeepti25/legal-ft-v0-updated") # Run inference sentences = [ "What new feature was introduced in ChatGPT's voice mode in December?", 'The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did.', 'The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 1024] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Information Retrieval * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.9167 | | cosine_accuracy@3 | 1.0 | | cosine_accuracy@5 | 1.0 | | cosine_accuracy@10 | 1.0 | | cosine_precision@1 | 0.9167 | | cosine_precision@3 | 0.3333 | | cosine_precision@5 | 0.2 | | cosine_precision@10 | 0.1 | | cosine_recall@1 | 0.9167 | | cosine_recall@3 | 1.0 | | cosine_recall@5 | 1.0 | | cosine_recall@10 | 1.0 | | **cosine_ndcg@10** | **0.9692** | | cosine_mrr@10 | 0.9583 | | cosine_map@100 | 0.9583 |   ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 156 training samples * Columns: <code>sentence_0</code> and <code>sentence_1</code> * Approximate statistics based on the first 156 samples: | | sentence_0 | sentence_1 | |:--------|:-----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| | type | string | string | | details | <ul><li>min: 13 tokens</li><li>mean: 20.17 tokens</li><li>max: 34 tokens</li></ul> | <ul><li>min: 43 tokens</li><li>mean: 135.18 tokens</li><li>max: 214 tokens</li></ul> | * Samples: | sentence_0 | sentence_1 | |:-------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | <code>What is the significance of prompt engineering in DALL-E 3?</code> | <code>Now add a walrus: Prompt engineering in DALL-E 3 32.8k 41.2k Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive 32.5k 38.2k ChatGPT can’t access the internet, even though it really looks like it can 30.5k 34.2k Stanford Alpaca, and the acceleration of on-device large language model development 29.7k 35.7k Run Llama 2 on your own Mac using LLM and Homebrew 27.9k 33.6k Midjourney 5.1 26.7k 33.4k Think of language models like ChatGPT as a “calculator for words” 25k 31.8k Multi-modal prompt injection image attacks against GPT-4V 23.7k 27.4k</code> | | <code>How does the vicuna-7b Large Language Model operate within a browser?</code> | <code>Now add a walrus: Prompt engineering in DALL-E 3 32.8k 41.2k Web LLM runs the vicuna-7b Large Language Model entirely in your browser, and it’s very impressive 32.5k 38.2k ChatGPT can’t access the internet, even though it really looks like it can 30.5k 34.2k Stanford Alpaca, and the acceleration of on-device large language model development 29.7k 35.7k Run Llama 2 on your own Mac using LLM and Homebrew 27.9k 33.6k Midjourney 5.1 26.7k 33.4k Think of language models like ChatGPT as a “calculator for words” 25k 31.8k Multi-modal prompt injection image attacks against GPT-4V 23.7k 27.4k</code> | | <code>What model of MacBook Pro is being used in the context, and what is its storage capacity?</code> | <code>My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment). That same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:</code> | * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 768, 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `num_train_epochs`: 10 - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `torch_empty_cache_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.0 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: False - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: None - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `include_for_metrics`: [] - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin </details> ### Training Logs | Epoch | Step | cosine_ndcg@10 | |:-----:|:----:|:--------------:| | 1.0 | 16 | 0.9692 | | 2.0 | 32 | 0.9692 | | 3.0 | 48 | 1.0 | | 3.125 | 50 | 1.0 | | 4.0 | 64 | 1.0 | | 5.0 | 80 | 0.9692 | | 6.0 | 96 | 0.9692 | | 6.25 | 100 | 0.9692 | | 7.0 | 112 | 0.9692 | | 8.0 | 128 | 0.9692 | | 9.0 | 144 | 0.9692 | | 9.375 | 150 | 0.9692 | | 10.0 | 160 | 0.9692 | ### Framework Versions - Python: 3.11.11 - Sentence Transformers: 3.4.1 - Transformers: 4.48.3 - PyTorch: 2.5.1+cu124 - Accelerate: 1.3.0 - Datasets: 3.3.1 - Tokenizers: 0.21.0 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

{"base_model": "Snowflake/snowflake-arctic-embed-l", "library_name": "sentence-transformers", "metrics": ["cosine_accuracy@1", "cosine_accuracy@3", "cosine_accuracy@5", "cosine_accuracy@10", "cosine_precision@1", "cosine_precision@3", "cosine_precision@5", "cosine_precision@10", "cosine_recall@1", "cosine_recall@3", "cosine_recall@5", "cosine_recall@10", "cosine_ndcg@10", "cosine_mrr@10", "cosine_map@100"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:156", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss"], "widget": [{"source_sentence": "What significant change occurred in the AI landscape regarding models surpassing GPT-4 in the past twelve months?", "sentences": ["Except... you can run generated code to see if it’s correct. And with patterns like ChatGPT Code Interpreter the LLM can execute the code itself, process the error message, then rewrite it and keep trying until it works!\nSo hallucination is a much lesser problem for code generation than for anything else. If only we had the equivalent of Code Interpreter for fact-checking natural language!\nHow should we feel about this as software engineers?\nOn the one hand, this feels like a threat: who needs a programmer if ChatGPT can write code for you?", "The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total.", "If you think about what they do, this isn’t such a big surprise. The grammar rules of programming languages like Python and JavaScript are massively less complicated than the grammar of Chinese, Spanish or English.\nIt’s still astonishing to me how effective they are though.\nOne of the great weaknesses of LLMs is their tendency to hallucinate—to imagine things that don’t correspond to reality. You would expect this to be a particularly bad problem for code—if an LLM hallucinates a method that doesn’t exist, the code should be useless."]}, {"source_sentence": "How does Claude enable users to interact with applications created through its interface?", "sentences": ["OpenAI made GPT-4o free for all users in May, and Claude 3.5 Sonnet was freely available from its launch in June. This was a momentus change, because for the previous year free users had mostly been restricted to GPT-3.5 level models, meaning new users got a very inaccurate mental model of what a capable LLM could actually do.\nThat era appears to have ended, likely permanently, with OpenAI’s launch of ChatGPT Pro. This $200/month subscription service is the only way to access their most capable model, o1 Pro.\nSince the trick behind the o1 series (and the future models it will undoubtedly inspire) is to expend more compute time to get better results, I don’t think those days of free access to the best available models are likely to return.", "We already knew LLMs were spookily good at writing code. If you prompt them right, it turns out they can build you a full interactive application using HTML, CSS and JavaScript (and tools like React if you wire up some extra supporting build mechanisms)—often in a single prompt.\nAnthropic kicked this idea into high gear when they released Claude Artifacts, a groundbreaking new feature that was initially slightly lost in the noise due to being described half way through their announcement of the incredible Claude 3.5 Sonnet.\nWith Artifacts, Claude can write you an on-demand interactive application and then let you use it directly inside the Claude interface.\nHere’s my Extract URLs app, entirely generated by Claude:", "Industry’s Tardy Response to the AI Prompt Injection Vulnerability on RedMonk Conversations\n\n\nPosted 31st December 2023 at 11:59 pm · Follow me on Mastodon or Twitter or subscribe to my newsletter\n\n\nMore recent articles\n\nLLM 0.22, the annotated release notes - 17th February 2025\nRun LLMs on macOS using llm-mlx and Apple's MLX framework - 15th February 2025\nURL-addressable Pyodide Python environments - 13th February 2025\n\n\n \n\n\nThis is Stuff we figured out about AI in 2023 by Simon Willison, posted on 31st December 2023.\n\nPart of series LLMs annual review\n\nStuff we figured out about AI in 2023 - Dec. 31, 2023, 11:59 p.m. \nThings we learned about LLMs in 2024 - Dec. 31, 2024, 6:07 p.m. \n\n\n\n blogging\n 69"]}, {"source_sentence": "What incident involving Google Search is mentioned in the context, and what was the nature of the misinformation?", "sentences": ["My personal laptop is a 64GB M2 MacBook Pro from 2023. It’s a powerful machine, but it’s also nearly two years old now—and crucially it’s the same laptop I’ve been using ever since I first ran an LLM on my computer back in March 2023 (see Large language models are having their Stable Diffusion moment).\nThat same laptop that could just about run a GPT-3-class model in March last year has now run multiple GPT-4 class models! Some of my notes on that:", "Terminology aside, I remain skeptical as to their utility based, once again, on the challenge of gullibility. LLMs believe anything you tell them. Any systems that attempts to make meaningful decisions on your behalf will run into the same roadblock: how good is a travel agent, or a digital assistant, or even a research tool if it can’t distinguish truth from fiction?\nJust the other day Google Search was caught serving up an entirely fake description of the non-existant movie “Encanto 2”. It turned out to be summarizing an imagined movie listing from a fan fiction wiki.", "On the other hand, as software engineers we are better placed to take advantage of this than anyone else. We’ve all been given weird coding interns—we can use our deep knowledge to prompt them to solve coding problems more effectively than anyone else can.\nThe ethics of this space remain diabolically complex\nIn September last year Andy Baio and I produced the first major story on the unlicensed training data behind Stable Diffusion.\nSince then, almost every major LLM (and most of the image generation models) have also been trained on unlicensed data."]}, {"source_sentence": "What are the limitations of Apple's LLM features compared to frontier LLMs, according to the context?", "sentences": ["DeepSeek v3 is a huge 685B parameter model—one of the largest openly licensed models currently available, significantly bigger than the largest of Meta’s Llama series, Llama 3.1 405B.\nBenchmarks put it up there with Claude 3.5 Sonnet. Vibe benchmarks (aka the Chatbot Arena) currently rank it 7th, just behind the Gemini 2.0 and OpenAI 4o/o1 models. This is by far the highest ranking openly licensed model.\nThe really impressive thing about DeepSeek v3 is the training cost. The model was trained on 2,788,000 H800 GPU hours at an estimated cost of $5,576,000. Llama 3.1 405B trained 30,840,000 GPU hours—11x that used by DeepSeek v3, for a model that benchmarks slightly worse.", "An interesting point of comparison here could be the way railways rolled out around the world in the 1800s. Constructing these required enormous investments and had a massive environmental impact, and many of the lines that were built turned out to be unnecessary—sometimes multiple lines from different companies serving the exact same routes!\nThe resulting bubbles contributed to several financial crashes, see Wikipedia for Panic of 1873, Panic of 1893, Panic of 1901 and the UK’s Railway Mania. They left us with a lot of useful infrastructure and a great deal of bankruptcies and environmental damage.\nThe year of slop", "Now that those features are rolling out they’re pretty weak. As an LLM power-user I know what these models are capable of, and Apple’s LLM features offer a pale imitation of what a frontier LLM can do. Instead we’re getting notification summaries that misrepresent news headlines and writing assistant tools that I’ve not found useful at all. Genmoji are kind of fun though.\nThe rise of inference-scaling “reasoning” models\nThe most interesting development in the final quarter of 2024 was the introduction of a new shape of LLM, exemplified by OpenAI’s o1 models—initially released as o1-preview and o1-mini on September 12th."]}, {"source_sentence": "What new feature was introduced in ChatGPT's voice mode in December?", "sentences": ["The most recent twist, again from December (December was a lot) is live video. ChatGPT voice mode now provides the option to share your camera feed with the model and talk about what you can see in real time. Google Gemini have a preview of the same feature, which they managed to ship the day before ChatGPT did.", "The two main categories I see are people who think AI agents are obviously things that go and act on your behalf—the travel agent model—and people who think in terms of LLMs that have been given access to tools which they can run in a loop as part of solving a problem. The term “autonomy” is often thrown into the mix too, again without including a clear definition.\n(I also collected 211 definitions on Twitter a few months ago—here they are in Datasette Lite—and had gemini-exp-1206 attempt to summarize them.)\nWhatever the term may mean, agents still have that feeling of perpetually “coming soon”.", "The GPT-4 barrier was comprehensively broken\nIn my December 2023 review I wrote about how We don’t yet know how to build GPT-4—OpenAI’s best model was almost a year old at that point, yet no other AI lab had produced anything better. What did OpenAI know that the rest of us didn’t?\nI’m relieved that this has changed completely in the past twelve months. 18 organizations now have models on the Chatbot Arena Leaderboard that rank higher than the original GPT-4 from March 2023 (GPT-4-0314 on the board)—70 models in total."]}], "model-index": [{"name": "SentenceTransformer based on Snowflake/snowflake-arctic-embed-l", "results": [{"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "Unknown", "type": "unknown"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.9166666666666666, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 1.0, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 1.0, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 1.0, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.9166666666666666, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.3333333333333333, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.20000000000000004, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.10000000000000002, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.9166666666666666, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 1.0, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 1.0, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 1.0, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.9692441461309548, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.9583333333333334, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.9583333333333334, "name": "Cosine Map@100"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,533

MoritzLaurer/deberta-v3-base-zeroshot-v2.0-c

MoritzLaurer

zero-shot-classification

[ "transformers", "safetensors", "deberta-v2", "text-classification", "zero-shot-classification", "en", "arxiv:2312.17543", "base_model:microsoft/deberta-v3-base", "base_model:finetune:microsoft/deberta-v3-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2024-03-21T22:21:55Z

2024-04-04T07:04:21+00:00

369

0

--- base_model: microsoft/deberta-v3-base language: - en library_name: transformers license: mit pipeline_tag: zero-shot-classification tags: - text-classification - zero-shot-classification --- # Model description: deberta-v3-base-zeroshot-v2.0-c ## zeroshot-v2.0 series of models Models in this series are designed for efficient zeroshot classification with the Hugging Face pipeline. These models can do classification without training data and run on both GPUs and CPUs. An overview of the latest zeroshot classifiers is available in my [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). The main update of this `zeroshot-v2.0` series of models is that several models are trained on fully commercially-friendly data for users with strict license requirements. These models can do one universal classification task: determine whether a hypothesis is "true" or "not true" given a text (`entailment` vs. `not_entailment`). This task format is based on the Natural Language Inference task (NLI). The task is so universal that any classification task can be reformulated into this task by the Hugging Face pipeline. ## Training data Models with a "`-c`" in the name are trained on two types of fully commercially-friendly data: 1. Synthetic data generated with [Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1). I first created a list of 500+ diverse text classification tasks for 25 professions in conversations with Mistral-large. The data was manually curated. I then used this as seed data to generate several hundred thousand texts for these tasks with Mixtral-8x7B-Instruct-v0.1. The final dataset used is available in the [synthetic_zeroshot_mixtral_v0.1](https://huggingface.co/datasets/MoritzLaurer/synthetic_zeroshot_mixtral_v0.1) dataset in the subset `mixtral_written_text_for_tasks_v4`. Data curation was done in multiple iterations and will be improved in future iterations. 2. Two commercially-friendly NLI datasets: ([MNLI](https://huggingface.co/datasets/nyu-mll/multi_nli), [FEVER-NLI](https://huggingface.co/datasets/fever)). These datasets were added to increase generalization. 3. Models without a "`-c`" in the name also included a broader mix of training data with a broader mix of licenses: ANLI, WANLI, LingNLI, and all datasets in [this list](https://github.com/MoritzLaurer/zeroshot-classifier/blob/7f82e4ab88d7aa82a4776f161b368cc9fa778001/v1_human_data/datasets_overview.csv) where `used_in_v1.1==True`. ## How to use the models ```python #!pip install transformers[sentencepiece] from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ``` `multi_label=False` forces the model to decide on only one class. `multi_label=True` enables the model to choose multiple classes. ## Metrics The models were evaluated on 28 different text classification tasks with the [f1_macro](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) metric. The main reference point is `facebook/bart-large-mnli` which is, at the time of writing (03.04.24), the most used commercially-friendly 0-shot classifier. ![results_aggreg_v2.0](https://raw.githubusercontent.com/MoritzLaurer/zeroshot-classifier/main/v2_synthetic_data/results/zeroshot-v2.0-aggreg.png) | | facebook/bart-large-mnli | roberta-base-zeroshot-v2.0-c | roberta-large-zeroshot-v2.0-c | deberta-v3-base-zeroshot-v2.0-c | deberta-v3-base-zeroshot-v2.0 (fewshot) | deberta-v3-large-zeroshot-v2.0-c | deberta-v3-large-zeroshot-v2.0 (fewshot) | bge-m3-zeroshot-v2.0-c | bge-m3-zeroshot-v2.0 (fewshot) | |:---------------------------|---------------------------:|-----------------------------:|------------------------------:|--------------------------------:|-----------------------------------:|---------------------------------:|------------------------------------:|-----------------------:|--------------------------:| | all datasets mean | 0.497 | 0.587 | 0.622 | 0.619 | 0.643 (0.834) | 0.676 | 0.673 (0.846) | 0.59 | (0.803) | | amazonpolarity (2) | 0.937 | 0.924 | 0.951 | 0.937 | 0.943 (0.961) | 0.952 | 0.956 (0.968) | 0.942 | (0.951) | | imdb (2) | 0.892 | 0.871 | 0.904 | 0.893 | 0.899 (0.936) | 0.923 | 0.918 (0.958) | 0.873 | (0.917) | | appreviews (2) | 0.934 | 0.913 | 0.937 | 0.938 | 0.945 (0.948) | 0.943 | 0.949 (0.962) | 0.932 | (0.954) | | yelpreviews (2) | 0.948 | 0.953 | 0.977 | 0.979 | 0.975 (0.989) | 0.988 | 0.985 (0.994) | 0.973 | (0.978) | | rottentomatoes (2) | 0.83 | 0.802 | 0.841 | 0.84 | 0.86 (0.902) | 0.869 | 0.868 (0.908) | 0.813 | (0.866) | | emotiondair (6) | 0.455 | 0.482 | 0.486 | 0.459 | 0.495 (0.748) | 0.499 | 0.484 (0.688) | 0.453 | (0.697) | | emocontext (4) | 0.497 | 0.555 | 0.63 | 0.59 | 0.592 (0.799) | 0.699 | 0.676 (0.81) | 0.61 | (0.798) | | empathetic (32) | 0.371 | 0.374 | 0.404 | 0.378 | 0.405 (0.53) | 0.447 | 0.478 (0.555) | 0.387 | (0.455) | | financialphrasebank (3) | 0.465 | 0.562 | 0.455 | 0.714 | 0.669 (0.906) | 0.691 | 0.582 (0.913) | 0.504 | (0.895) | | banking77 (72) | 0.312 | 0.124 | 0.29 | 0.421 | 0.446 (0.751) | 0.513 | 0.567 (0.766) | 0.387 | (0.715) | | massive (59) | 0.43 | 0.428 | 0.543 | 0.512 | 0.52 (0.755) | 0.526 | 0.518 (0.789) | 0.414 | (0.692) | | wikitoxic_toxicaggreg (2) | 0.547 | 0.751 | 0.766 | 0.751 | 0.769 (0.904) | 0.741 | 0.787 (0.911) | 0.736 | (0.9) | | wikitoxic_obscene (2) | 0.713 | 0.817 | 0.854 | 0.853 | 0.869 (0.922) | 0.883 | 0.893 (0.933) | 0.783 | (0.914) | | wikitoxic_threat (2) | 0.295 | 0.71 | 0.817 | 0.813 | 0.87 (0.946) | 0.827 | 0.879 (0.952) | 0.68 | (0.947) | | wikitoxic_insult (2) | 0.372 | 0.724 | 0.798 | 0.759 | 0.811 (0.912) | 0.77 | 0.779 (0.924) | 0.783 | (0.915) | | wikitoxic_identityhate (2) | 0.473 | 0.774 | 0.798 | 0.774 | 0.765 (0.938) | 0.797 | 0.806 (0.948) | 0.761 | (0.931) | | hateoffensive (3) | 0.161 | 0.352 | 0.29 | 0.315 | 0.371 (0.862) | 0.47 | 0.461 (0.847) | 0.291 | (0.823) | | hatexplain (3) | 0.239 | 0.396 | 0.314 | 0.376 | 0.369 (0.765) | 0.378 | 0.389 (0.764) | 0.29 | (0.729) | | biasframes_offensive (2) | 0.336 | 0.571 | 0.583 | 0.544 | 0.601 (0.867) | 0.644 | 0.656 (0.883) | 0.541 | (0.855) | | biasframes_sex (2) | 0.263 | 0.617 | 0.835 | 0.741 | 0.809 (0.922) | 0.846 | 0.815 (0.946) | 0.748 | (0.905) | | biasframes_intent (2) | 0.616 | 0.531 | 0.635 | 0.554 | 0.61 (0.881) | 0.696 | 0.687 (0.891) | 0.467 | (0.868) | | agnews (4) | 0.703 | 0.758 | 0.745 | 0.68 | 0.742 (0.898) | 0.819 | 0.771 (0.898) | 0.687 | (0.892) | | yahootopics (10) | 0.299 | 0.543 | 0.62 | 0.578 | 0.564 (0.722) | 0.621 | 0.613 (0.738) | 0.587 | (0.711) | | trueteacher (2) | 0.491 | 0.469 | 0.402 | 0.431 | 0.479 (0.82) | 0.459 | 0.538 (0.846) | 0.471 | (0.518) | | spam (2) | 0.505 | 0.528 | 0.504 | 0.507 | 0.464 (0.973) | 0.74 | 0.597 (0.983) | 0.441 | (0.978) | | wellformedquery (2) | 0.407 | 0.333 | 0.333 | 0.335 | 0.491 (0.769) | 0.334 | 0.429 (0.815) | 0.361 | (0.718) | | manifesto (56) | 0.084 | 0.102 | 0.182 | 0.17 | 0.187 (0.376) | 0.258 | 0.256 (0.408) | 0.147 | (0.331) | | capsotu (21) | 0.34 | 0.479 | 0.523 | 0.502 | 0.477 (0.664) | 0.603 | 0.502 (0.686) | 0.472 | (0.644) | These numbers indicate zeroshot performance, as no data from these datasets was added in the training mix. Note that models without a "`-c`" in the title were evaluated twice: one run without any data from these 28 datasets to test pure zeroshot performance (the first number in the respective column) and the final run including up to 500 training data points per class from each of the 28 datasets (the second number in brackets in the column, "fewshot"). No model was trained on test data. Details on the different datasets are available here: https://github.com/MoritzLaurer/zeroshot-classifier/blob/main/v1_human_data/datasets_overview.csv ## When to use which model - **deberta-v3-zeroshot vs. roberta-zeroshot**: deberta-v3 performs clearly better than roberta, but it is a bit slower. roberta is directly compatible with Hugging Face's production inference TEI containers and flash attention. These containers are a good choice for production use-cases. tl;dr: For accuracy, use a deberta-v3 model. If production inference speed is a concern, you can consider a roberta model (e.g. in a TEI container and [HF Inference Endpoints](https://ui.endpoints.huggingface.co/catalog)). - **commercial use-cases**: models with "`-c`" in the title are guaranteed to be trained on only commercially-friendly data. Models without a "`-c`" were trained on more data and perform better, but include data with non-commercial licenses. Legal opinions diverge if this training data affects the license of the trained model. For users with strict legal requirements, the models with "`-c`" in the title are recommended. - **Multilingual/non-English use-cases**: use [bge-m3-zeroshot-v2.0](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0) or [bge-m3-zeroshot-v2.0-c](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0-c). Note that multilingual models perform worse than English-only models. You can therefore also first machine translate your texts to English with libraries like [EasyNMT](https://github.com/UKPLab/EasyNMT) and then apply any English-only model to the translated data. Machine translation also facilitates validation in case your team does not speak all languages in the data. - **context window**: The `bge-m3` models can process up to 8192 tokens. The other models can process up to 512. Note that longer text inputs both make the mode slower and decrease performance, so if you're only working with texts of up to 400~ words / 1 page, use e.g. a deberta model for better performance. - The latest updates on new models are always available in the [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). ## Reproduction Reproduction code is available in the `v2_synthetic_data` directory here: https://github.com/MoritzLaurer/zeroshot-classifier/tree/main ## Limitations and bias The model can only do text classification tasks. Biases can come from the underlying foundation model, the human NLI training data and the synthetic data generated by Mixtral. ## License The foundation model was published under the MIT license. The licenses of the training data vary depending on the model, see above. ## Citation This model is an extension of the research described in this [paper](https://arxiv.org/pdf/2312.17543.pdf). If you use this model academically, please cite: ``` @misc{laurer_building_2023, title = {Building {Efficient} {Universal} {Classifiers} with {Natural} {Language} {Inference}}, url = {http://arxiv.org/abs/2312.17543}, doi = {10.48550/arXiv.2312.17543}, abstract = {Generative Large Language Models (LLMs) have become the mainstream choice for fewshot and zeroshot learning thanks to the universality of text generation. Many users, however, do not need the broad capabilities of generative LLMs when they only want to automate a classification task. Smaller BERT-like models can also learn universal tasks, which allow them to do any text classification task without requiring fine-tuning (zeroshot classification) or to learn new tasks with only a few examples (fewshot), while being significantly more efficient than generative LLMs. This paper (1) explains how Natural Language Inference (NLI) can be used as a universal classification task that follows similar principles as instruction fine-tuning of generative LLMs, (2) provides a step-by-step guide with reusable Jupyter notebooks for building a universal classifier, and (3) shares the resulting universal classifier that is trained on 33 datasets with 389 diverse classes. Parts of the code we share has been used to train our older zeroshot classifiers that have been downloaded more than 55 million times via the Hugging Face Hub as of December 2023. Our new classifier improves zeroshot performance by 9.4\%.}, urldate = {2024-01-05}, publisher = {arXiv}, author = {Laurer, Moritz and van Atteveldt, Wouter and Casas, Andreu and Welbers, Kasper}, month = dec, year = {2023}, note = {arXiv:2312.17543 [cs]}, keywords = {Computer Science - Artificial Intelligence, Computer Science - Computation and Language}, } ``` ### Ideas for cooperation or questions? If you have questions or ideas for cooperation, contact me at moritz{at}huggingface{dot}co or [LinkedIn](https://www.linkedin.com/in/moritz-laurer/) ### Flexible usage and "prompting" You can formulate your own hypotheses by changing the `hypothesis_template` of the zeroshot pipeline. Similar to "prompt engineering" for LLMs, you can test different formulations of your `hypothesis_template` and verbalized classes to improve performance. ```python from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" # formulation 1 hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] # formulation 2 depending on your use-case hypothesis_template = "The topic of this text is {}" classes_verbalized = ["political activities", "economic policy", "entertainment or music", "environmental protection"] # test different formulations zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ```

null

Non_BioNLP

# Model description: deberta-v3-base-zeroshot-v2.0-c ## zeroshot-v2.0 series of models Models in this series are designed for efficient zeroshot classification with the Hugging Face pipeline. These models can do classification without training data and run on both GPUs and CPUs. An overview of the latest zeroshot classifiers is available in my [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). The main update of this `zeroshot-v2.0` series of models is that several models are trained on fully commercially-friendly data for users with strict license requirements. These models can do one universal classification task: determine whether a hypothesis is "true" or "not true" given a text (`entailment` vs. `not_entailment`). This task format is based on the Natural Language Inference task (NLI). The task is so universal that any classification task can be reformulated into this task by the Hugging Face pipeline. ## Training data Models with a "`-c`" in the name are trained on two types of fully commercially-friendly data: 1. Synthetic data generated with [Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1). I first created a list of 500+ diverse text classification tasks for 25 professions in conversations with Mistral-large. The data was manually curated. I then used this as seed data to generate several hundred thousand texts for these tasks with Mixtral-8x7B-Instruct-v0.1. The final dataset used is available in the [synthetic_zeroshot_mixtral_v0.1](https://huggingface.co/datasets/MoritzLaurer/synthetic_zeroshot_mixtral_v0.1) dataset in the subset `mixtral_written_text_for_tasks_v4`. Data curation was done in multiple iterations and will be improved in future iterations. 2. Two commercially-friendly NLI datasets: ([MNLI](https://huggingface.co/datasets/nyu-mll/multi_nli), [FEVER-NLI](https://huggingface.co/datasets/fever)). These datasets were added to increase generalization. 3. Models without a "`-c`" in the name also included a broader mix of training data with a broader mix of licenses: ANLI, WANLI, LingNLI, and all datasets in [this list](https://github.com/MoritzLaurer/zeroshot-classifier/blob/7f82e4ab88d7aa82a4776f161b368cc9fa778001/v1_human_data/datasets_overview.csv) where `used_in_v1.1==True`. ## How to use the models ```python #!pip install transformers[sentencepiece] from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ``` `multi_label=False` forces the model to decide on only one class. `multi_label=True` enables the model to choose multiple classes. ## Metrics The models were evaluated on 28 different text classification tasks with the [f1_macro](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) metric. The main reference point is `facebook/bart-large-mnli` which is, at the time of writing (03.04.24), the most used commercially-friendly 0-shot classifier. ![results_aggreg_v2.0](https://raw.githubusercontent.com/MoritzLaurer/zeroshot-classifier/main/v2_synthetic_data/results/zeroshot-v2.0-aggreg.png) | | facebook/bart-large-mnli | roberta-base-zeroshot-v2.0-c | roberta-large-zeroshot-v2.0-c | deberta-v3-base-zeroshot-v2.0-c | deberta-v3-base-zeroshot-v2.0 (fewshot) | deberta-v3-large-zeroshot-v2.0-c | deberta-v3-large-zeroshot-v2.0 (fewshot) | bge-m3-zeroshot-v2.0-c | bge-m3-zeroshot-v2.0 (fewshot) | |:---------------------------|---------------------------:|-----------------------------:|------------------------------:|--------------------------------:|-----------------------------------:|---------------------------------:|------------------------------------:|-----------------------:|--------------------------:| | all datasets mean | 0.497 | 0.587 | 0.622 | 0.619 | 0.643 (0.834) | 0.676 | 0.673 (0.846) | 0.59 | (0.803) | | amazonpolarity (2) | 0.937 | 0.924 | 0.951 | 0.937 | 0.943 (0.961) | 0.952 | 0.956 (0.968) | 0.942 | (0.951) | | imdb (2) | 0.892 | 0.871 | 0.904 | 0.893 | 0.899 (0.936) | 0.923 | 0.918 (0.958) | 0.873 | (0.917) | | appreviews (2) | 0.934 | 0.913 | 0.937 | 0.938 | 0.945 (0.948) | 0.943 | 0.949 (0.962) | 0.932 | (0.954) | | yelpreviews (2) | 0.948 | 0.953 | 0.977 | 0.979 | 0.975 (0.989) | 0.988 | 0.985 (0.994) | 0.973 | (0.978) | | rottentomatoes (2) | 0.83 | 0.802 | 0.841 | 0.84 | 0.86 (0.902) | 0.869 | 0.868 (0.908) | 0.813 | (0.866) | | emotiondair (6) | 0.455 | 0.482 | 0.486 | 0.459 | 0.495 (0.748) | 0.499 | 0.484 (0.688) | 0.453 | (0.697) | | emocontext (4) | 0.497 | 0.555 | 0.63 | 0.59 | 0.592 (0.799) | 0.699 | 0.676 (0.81) | 0.61 | (0.798) | | empathetic (32) | 0.371 | 0.374 | 0.404 | 0.378 | 0.405 (0.53) | 0.447 | 0.478 (0.555) | 0.387 | (0.455) | | financialphrasebank (3) | 0.465 | 0.562 | 0.455 | 0.714 | 0.669 (0.906) | 0.691 | 0.582 (0.913) | 0.504 | (0.895) | | banking77 (72) | 0.312 | 0.124 | 0.29 | 0.421 | 0.446 (0.751) | 0.513 | 0.567 (0.766) | 0.387 | (0.715) | | massive (59) | 0.43 | 0.428 | 0.543 | 0.512 | 0.52 (0.755) | 0.526 | 0.518 (0.789) | 0.414 | (0.692) | | wikitoxic_toxicaggreg (2) | 0.547 | 0.751 | 0.766 | 0.751 | 0.769 (0.904) | 0.741 | 0.787 (0.911) | 0.736 | (0.9) | | wikitoxic_obscene (2) | 0.713 | 0.817 | 0.854 | 0.853 | 0.869 (0.922) | 0.883 | 0.893 (0.933) | 0.783 | (0.914) | | wikitoxic_threat (2) | 0.295 | 0.71 | 0.817 | 0.813 | 0.87 (0.946) | 0.827 | 0.879 (0.952) | 0.68 | (0.947) | | wikitoxic_insult (2) | 0.372 | 0.724 | 0.798 | 0.759 | 0.811 (0.912) | 0.77 | 0.779 (0.924) | 0.783 | (0.915) | | wikitoxic_identityhate (2) | 0.473 | 0.774 | 0.798 | 0.774 | 0.765 (0.938) | 0.797 | 0.806 (0.948) | 0.761 | (0.931) | | hateoffensive (3) | 0.161 | 0.352 | 0.29 | 0.315 | 0.371 (0.862) | 0.47 | 0.461 (0.847) | 0.291 | (0.823) | | hatexplain (3) | 0.239 | 0.396 | 0.314 | 0.376 | 0.369 (0.765) | 0.378 | 0.389 (0.764) | 0.29 | (0.729) | | biasframes_offensive (2) | 0.336 | 0.571 | 0.583 | 0.544 | 0.601 (0.867) | 0.644 | 0.656 (0.883) | 0.541 | (0.855) | | biasframes_sex (2) | 0.263 | 0.617 | 0.835 | 0.741 | 0.809 (0.922) | 0.846 | 0.815 (0.946) | 0.748 | (0.905) | | biasframes_intent (2) | 0.616 | 0.531 | 0.635 | 0.554 | 0.61 (0.881) | 0.696 | 0.687 (0.891) | 0.467 | (0.868) | | agnews (4) | 0.703 | 0.758 | 0.745 | 0.68 | 0.742 (0.898) | 0.819 | 0.771 (0.898) | 0.687 | (0.892) | | yahootopics (10) | 0.299 | 0.543 | 0.62 | 0.578 | 0.564 (0.722) | 0.621 | 0.613 (0.738) | 0.587 | (0.711) | | trueteacher (2) | 0.491 | 0.469 | 0.402 | 0.431 | 0.479 (0.82) | 0.459 | 0.538 (0.846) | 0.471 | (0.518) | | spam (2) | 0.505 | 0.528 | 0.504 | 0.507 | 0.464 (0.973) | 0.74 | 0.597 (0.983) | 0.441 | (0.978) | | wellformedquery (2) | 0.407 | 0.333 | 0.333 | 0.335 | 0.491 (0.769) | 0.334 | 0.429 (0.815) | 0.361 | (0.718) | | manifesto (56) | 0.084 | 0.102 | 0.182 | 0.17 | 0.187 (0.376) | 0.258 | 0.256 (0.408) | 0.147 | (0.331) | | capsotu (21) | 0.34 | 0.479 | 0.523 | 0.502 | 0.477 (0.664) | 0.603 | 0.502 (0.686) | 0.472 | (0.644) | These numbers indicate zeroshot performance, as no data from these datasets was added in the training mix. Note that models without a "`-c`" in the title were evaluated twice: one run without any data from these 28 datasets to test pure zeroshot performance (the first number in the respective column) and the final run including up to 500 training data points per class from each of the 28 datasets (the second number in brackets in the column, "fewshot"). No model was trained on test data. Details on the different datasets are available here: https://github.com/MoritzLaurer/zeroshot-classifier/blob/main/v1_human_data/datasets_overview.csv ## When to use which model - **deberta-v3-zeroshot vs. roberta-zeroshot**: deberta-v3 performs clearly better than roberta, but it is a bit slower. roberta is directly compatible with Hugging Face's production inference TEI containers and flash attention. These containers are a good choice for production use-cases. tl;dr: For accuracy, use a deberta-v3 model. If production inference speed is a concern, you can consider a roberta model (e.g. in a TEI container and [HF Inference Endpoints](https://ui.endpoints.huggingface.co/catalog)). - **commercial use-cases**: models with "`-c`" in the title are guaranteed to be trained on only commercially-friendly data. Models without a "`-c`" were trained on more data and perform better, but include data with non-commercial licenses. Legal opinions diverge if this training data affects the license of the trained model. For users with strict legal requirements, the models with "`-c`" in the title are recommended. - **Multilingual/non-English use-cases**: use [bge-m3-zeroshot-v2.0](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0) or [bge-m3-zeroshot-v2.0-c](https://huggingface.co/MoritzLaurer/bge-m3-zeroshot-v2.0-c). Note that multilingual models perform worse than English-only models. You can therefore also first machine translate your texts to English with libraries like [EasyNMT](https://github.com/UKPLab/EasyNMT) and then apply any English-only model to the translated data. Machine translation also facilitates validation in case your team does not speak all languages in the data. - **context window**: The `bge-m3` models can process up to 8192 tokens. The other models can process up to 512. Note that longer text inputs both make the mode slower and decrease performance, so if you're only working with texts of up to 400~ words / 1 page, use e.g. a deberta model for better performance. - The latest updates on new models are always available in the [Zeroshot Classifier Collection](https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f). ## Reproduction Reproduction code is available in the `v2_synthetic_data` directory here: https://github.com/MoritzLaurer/zeroshot-classifier/tree/main ## Limitations and bias The model can only do text classification tasks. Biases can come from the underlying foundation model, the human NLI training data and the synthetic data generated by Mixtral. ## License The foundation model was published under the MIT license. The licenses of the training data vary depending on the model, see above. ## Citation This model is an extension of the research described in this [paper](https://arxiv.org/pdf/2312.17543.pdf). If you use this model academically, please cite: ``` @misc{laurer_building_2023, title = {Building {Efficient} {Universal} {Classifiers} with {Natural} {Language} {Inference}}, url = {http://arxiv.org/abs/2312.17543}, doi = {10.48550/arXiv.2312.17543}, abstract = {Generative Large Language Models (LLMs) have become the mainstream choice for fewshot and zeroshot learning thanks to the universality of text generation. Many users, however, do not need the broad capabilities of generative LLMs when they only want to automate a classification task. Smaller BERT-like models can also learn universal tasks, which allow them to do any text classification task without requiring fine-tuning (zeroshot classification) or to learn new tasks with only a few examples (fewshot), while being significantly more efficient than generative LLMs. This paper (1) explains how Natural Language Inference (NLI) can be used as a universal classification task that follows similar principles as instruction fine-tuning of generative LLMs, (2) provides a step-by-step guide with reusable Jupyter notebooks for building a universal classifier, and (3) shares the resulting universal classifier that is trained on 33 datasets with 389 diverse classes. Parts of the code we share has been used to train our older zeroshot classifiers that have been downloaded more than 55 million times via the Hugging Face Hub as of December 2023. Our new classifier improves zeroshot performance by 9.4\%.}, urldate = {2024-01-05}, publisher = {arXiv}, author = {Laurer, Moritz and van Atteveldt, Wouter and Casas, Andreu and Welbers, Kasper}, month = dec, year = {2023}, note = {arXiv:2312.17543 [cs]}, keywords = {Computer Science - Artificial Intelligence, Computer Science - Computation and Language}, } ``` ### Ideas for cooperation or questions? If you have questions or ideas for cooperation, contact me at moritz{at}huggingface{dot}co or [LinkedIn](https://www.linkedin.com/in/moritz-laurer/) ### Flexible usage and "prompting" You can formulate your own hypotheses by changing the `hypothesis_template` of the zeroshot pipeline. Similar to "prompt engineering" for LLMs, you can test different formulations of your `hypothesis_template` and verbalized classes to improve performance. ```python from transformers import pipeline text = "Angela Merkel is a politician in Germany and leader of the CDU" # formulation 1 hypothesis_template = "This text is about {}" classes_verbalized = ["politics", "economy", "entertainment", "environment"] # formulation 2 depending on your use-case hypothesis_template = "The topic of this text is {}" classes_verbalized = ["political activities", "economic policy", "entertainment or music", "environmental protection"] # test different formulations zeroshot_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/deberta-v3-large-zeroshot-v2.0") # change the model identifier here output = zeroshot_classifier(text, classes_verbalized, hypothesis_template=hypothesis_template, multi_label=False) print(output) ```

{"base_model": "microsoft/deberta-v3-base", "language": ["en"], "library_name": "transformers", "license": "mit", "pipeline_tag": "zero-shot-classification", "tags": ["text-classification", "zero-shot-classification"]}

task

[ "TEXT_CLASSIFICATION", "TRANSLATION" ]

44,534

cstr/whisper-large-v3-turbo-int8_float32

cstr

automatic-speech-recognition

[ "audio", "automatic-speech-recognition", "hf-asr-leaderboard", "en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", "no", "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su", "arxiv:2212.04356", "license:apache-2.0", "region:us" ]

2024-10-02T06:17:42Z

2024-10-02T07:23:38+00:00

57

0

--- language: - en - zh - de - es - ru - ko - fr - ja - pt - tr - pl - ca - nl - ar - sv - it - id - hi - fi - vi - he - uk - el - ms - cs - ro - da - hu - ta - false - th - ur - hr - bg - lt - la - mi - ml - cy - sk - te - fa - lv - bn - sr - az - sl - kn - et - mk - br - eu - is - hy - ne - mn - bs - kk - sq - sw - gl - mr - pa - si - km - sn - yo - so - af - oc - ka - be - tg - sd - gu - am - yi - lo - uz - fo - ht - ps - tk - nn - mt - sa - lb - my - bo - tl - mg - as - tt - haw - ln - ha - ba - jw - su license: apache-2.0 pipeline_tag: automatic-speech-recognition tags: - audio - automatic-speech-recognition - hf-asr-leaderboard widget: - example_title: Librispeech sample 1 src: https://cdn-media.huggingface.co/speech_samples/sample1.flac - example_title: Librispeech sample 2 src: https://cdn-media.huggingface.co/speech_samples/sample2.flac --- # Whisper turbo v3 for CTranslate2/faster-whisper This is an int8 quantified version of the OpenAI Whisper v3 turbo model. You can use it like this (example for usage in kaggle/Colab): ```kaggle # Clone the repository !git clone https://github.com/SYSTRAN/faster-whisper.git %cd faster-whisper # Install requirements !pip install -r requirements.txt # Import necessary modules from faster_whisper import WhisperModel from faster_whisper.transcribe import BatchedInferencePipeline # Initialize the model model = WhisperModel("cstr/whisper-large-v3-turbo-int8_float32", device="auto", compute_type="int8") batched_model = BatchedInferencePipeline(model=model) # Change to home directory and download audio %cd ~ !wget -c https://mcdn.podbean.com/mf/web/dir5wty678b6g4vg/HoP_453_-_The_Price_is_Right_-_Law_and_Economics_in_the_Second_Scholastic5yxzh.mp3 -O audio.mp3 # Benchmark transcription time import time import osstart_time = time.time() segments, info = batched_model.transcribe("audio.mp3", batch_size=16) end_time = time.time() # Print transcription for segment in segments: print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text)) # Print transcription info and benchmark results print(f"\nLanguage: {info.language}, Probability: {info.language_probability:.2f}") print(f"Duration: {info.duration:.2f}s, Duration after VAD: {info.duration_after_vad:.2f}s") transcription_time = end_time - start_time print(f"\nTranscription time: {transcription_time:.2f} seconds") # Calculate real-time factor real_time_factor = info.duration / transcription_time print(f"Real-time factor: {real_time_factor:.2f}x") # Print audio file size audio_file_size = os.path.getsize("audio.mp3") / (1024 * 1024) # Size in MB print(f"Audio file size: {audio_file_size:.2f} MB") ``` # Whisper (original Model Card follows) Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://huggingface.co/papers/2212.04356) by Alec Radford et al. from OpenAI. Trained on >5M hours of labeled data, Whisper demonstrates a strong ability to generalise to many datasets and domains in a zero-shot setting. Whisper large-v3-turbo is a distilled version of [Whisper large-v3](https://huggingface.co/openai/whisper-large-v3). In other words, it's the exact same model, except that the number of decoding layers have reduced from 32 to 4. As a result, the model is way faster, at the expense of a minor quality degradation. **Disclaimer**: Content for this model card has partly been written by the 🤗 Hugging Face team, and partly copied and pasted from the original model card. ## Usage Whisper large-v3-turbo is supported in Hugging Face 🤗 Transformers. To run the model, first install the Transformers library. For this example, we'll also install 🤗 Datasets to load toy audio dataset from the Hugging Face Hub, and 🤗 Accelerate to reduce the model loading time: ```bash pip install --upgrade pip pip install --upgrade transformers datasets[audio] accelerate ``` The model can be used with the [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) class to transcribe audios of arbitrary length: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` To transcribe a local audio file, simply pass the path to your audio file when you call the pipeline: ```python result = pipe("audio.mp3") ``` Multiple audio files can be transcribed in parallel by specifying them as a list and setting the `batch_size` parameter: ```python result = pipe(["audio_1.mp3", "audio_2.mp3"], batch_size=2) ``` Transformers is compatible with all Whisper decoding strategies, such as temperature fallback and condition on previous tokens. The following example demonstrates how to enable these heuristics: ```python generate_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } result = pipe(sample, generate_kwargs=generate_kwargs) ``` Whisper predicts the language of the source audio automatically. If the source audio language is known *a-priori*, it can be passed as an argument to the pipeline: ```python result = pipe(sample, generate_kwargs={"language": "english"}) ``` By default, Whisper performs the task of *speech transcription*, where the source audio language is the same as the target text language. To perform *speech translation*, where the target text is in English, set the task to `"translate"`: ```python result = pipe(sample, generate_kwargs={"task": "translate"}) ``` Finally, the model can be made to predict timestamps. For sentence-level timestamps, pass the `return_timestamps` argument: ```python result = pipe(sample, return_timestamps=True) print(result["chunks"]) ``` And for word-level timestamps: ```python result = pipe(sample, return_timestamps="word") print(result["chunks"]) ``` The above arguments can be used in isolation or in combination. For example, to perform the task of speech transcription where the source audio is in French, and we want to return sentence-level timestamps, the following can be used: ```python result = pipe(sample, return_timestamps=True, generate_kwargs={"language": "french", "task": "translate"}) print(result["chunks"]) ``` <details> <summary> For more control over the generation parameters, use the model + processor API directly: </summary> ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor from datasets import Audio, load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") dataset = dataset.cast_column("audio", Audio(processor.feature_extractor.sampling_rate)) sample = dataset[0]["audio"] inputs = processor( sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt", truncation=False, padding="longest", return_attention_mask=True, ) inputs = inputs.to(device, dtype=torch_dtype) gen_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } pred_ids = model.generate(**inputs, **gen_kwargs) pred_text = processor.batch_decode(pred_ids, skip_special_tokens=True, decode_with_timestamps=False) print(pred_text) ``` </details> ## Additional Speed & Memory Improvements You can apply additional speed and memory improvements to Whisper to further reduce the inference speed and VRAM requirements. ### Chunked Long-Form Whisper has a receptive field of 30-seconds. To transcribe audios longer than this, one of two long-form algorithms are required: 1. **Sequential:** uses a "sliding window" for buffered inference, transcribing 30-second slices one after the other 2. **Chunked:** splits long audio files into shorter ones (with a small overlap between segments), transcribes each segment independently, and stitches the resulting transcriptions at the boundaries The sequential long-form algorithm should be used in either of the following scenarios: 1. Transcription accuracy is the most important factor, and speed is less of a consideration 2. You are transcribing **batches** of long audio files, in which case the latency of sequential is comparable to chunked, while being up to 0.5% WER more accurate Conversely, the chunked algorithm should be used when: 1. Transcription speed is the most important factor 2. You are transcribing a **single** long audio file By default, Transformers uses the sequential algorithm. To enable the chunked algorithm, pass the `chunk_length_s` parameter to the `pipeline`. For large-v3, a chunk length of 30-seconds is optimal. To activate batching over long audio files, pass the argument `batch_size`: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, chunk_length_s=30, batch_size=16, # batch size for inference - set based on your device torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` #### Torch compile The Whisper forward pass is compatible with [`torch.compile`](https://pytorch.org/docs/stable/generated/torch.compile.html) for 4.5x speed-ups. **Note:** `torch.compile` is currently not compatible with the Chunked long-form algorithm or Flash Attention 2 ⚠️ ```python import torch from torch.nn.attention import SDPBackend, sdpa_kernel from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset from tqdm import tqdm torch.set_float32_matmul_precision("high") device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ).to(device) # Enable static cache and compile the forward pass model.generation_config.cache_implementation = "static" model.generation_config.max_new_tokens = 256 model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] # 2 warmup steps for _ in tqdm(range(2), desc="Warm-up step"): with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy(), generate_kwargs={"min_new_tokens": 256, "max_new_tokens": 256}) # fast run with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy()) print(result["text"]) ``` #### Flash Attention 2 We recommend using [Flash-Attention 2](https://huggingface.co/docs/transformers/main/en/perf_infer_gpu_one#flashattention-2) if your GPU supports it and you are not using [torch.compile](#torch-compile). To do so, first install [Flash Attention](https://github.com/Dao-AILab/flash-attention): ``` pip install flash-attn --no-build-isolation ``` Then pass `attn_implementation="flash_attention_2"` to `from_pretrained`: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="flash_attention_2") ``` #### Torch Scale-Product-Attention (SDPA) If your GPU does not support Flash Attention, we recommend making use of PyTorch [scaled dot-product attention (SDPA)](https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html). This attention implementation is activated **by default** for PyTorch versions 2.1.1 or greater. To check whether you have a compatible PyTorch version, run the following Python code snippet: ```python from transformers.utils import is_torch_sdpa_available print(is_torch_sdpa_available()) ``` If the above returns `True`, you have a valid version of PyTorch installed and SDPA is activated by default. If it returns `False`, you need to upgrade your PyTorch version according to the [official instructions](https://pytorch.org/get-started/locally/) Once a valid PyTorch version is installed, SDPA is activated by default. It can also be set explicitly by specifying `attn_implementation="sdpa"` as follows: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="sdpa") ``` For more information about how to use the SDPA refer to the [Transformers SDPA documentation](https://huggingface.co/docs/transformers/en/perf_infer_gpu_one#pytorch-scaled-dot-product-attention). ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. There are two flavours of Whisper model: English-only and multilingual. The English-only models were trained on the task of English speech recognition. The multilingual models were trained simultaneously on multilingual speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the *same* language as the audio. For speech translation, the model predicts transcriptions to a *different* language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are available as English-only and multilingual. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: | Size | Parameters | English-only | Multilingual | |----------|------------|------------------------------------------------------|-----------------------------------------------------| | tiny | 39 M | [✓](https://huggingface.co/openai/whisper-tiny.en) | [✓](https://huggingface.co/openai/whisper-tiny) | | base | 74 M | [✓](https://huggingface.co/openai/whisper-base.en) | [✓](https://huggingface.co/openai/whisper-base) | | small | 244 M | [✓](https://huggingface.co/openai/whisper-small.en) | [✓](https://huggingface.co/openai/whisper-small) | | medium | 769 M | [✓](https://huggingface.co/openai/whisper-medium.en) | [✓](https://huggingface.co/openai/whisper-medium) | | large | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large) | | large-v2 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v2) | | large-v3 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v3) | | large-v3-turbo | 809 M | x | [✓](https://huggingface.co/ylacombe/whisper-large-v3-turbo) | ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through *fine-tuning*. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data No information provided. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```

null

Non_BioNLP

# Whisper turbo v3 for CTranslate2/faster-whisper This is an int8 quantified version of the OpenAI Whisper v3 turbo model. You can use it like this (example for usage in kaggle/Colab): ```kaggle # Clone the repository !git clone https://github.com/SYSTRAN/faster-whisper.git %cd faster-whisper # Install requirements !pip install -r requirements.txt # Import necessary modules from faster_whisper import WhisperModel from faster_whisper.transcribe import BatchedInferencePipeline # Initialize the model model = WhisperModel("cstr/whisper-large-v3-turbo-int8_float32", device="auto", compute_type="int8") batched_model = BatchedInferencePipeline(model=model) # Change to home directory and download audio %cd ~ !wget -c https://mcdn.podbean.com/mf/web/dir5wty678b6g4vg/HoP_453_-_The_Price_is_Right_-_Law_and_Economics_in_the_Second_Scholastic5yxzh.mp3 -O audio.mp3 # Benchmark transcription time import time import osstart_time = time.time() segments, info = batched_model.transcribe("audio.mp3", batch_size=16) end_time = time.time() # Print transcription for segment in segments: print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text)) # Print transcription info and benchmark results print(f"\nLanguage: {info.language}, Probability: {info.language_probability:.2f}") print(f"Duration: {info.duration:.2f}s, Duration after VAD: {info.duration_after_vad:.2f}s") transcription_time = end_time - start_time print(f"\nTranscription time: {transcription_time:.2f} seconds") # Calculate real-time factor real_time_factor = info.duration / transcription_time print(f"Real-time factor: {real_time_factor:.2f}x") # Print audio file size audio_file_size = os.path.getsize("audio.mp3") / (1024 * 1024) # Size in MB print(f"Audio file size: {audio_file_size:.2f} MB") ``` # Whisper (original Model Card follows) Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://huggingface.co/papers/2212.04356) by Alec Radford et al. from OpenAI. Trained on >5M hours of labeled data, Whisper demonstrates a strong ability to generalise to many datasets and domains in a zero-shot setting. Whisper large-v3-turbo is a distilled version of [Whisper large-v3](https://huggingface.co/openai/whisper-large-v3). In other words, it's the exact same model, except that the number of decoding layers have reduced from 32 to 4. As a result, the model is way faster, at the expense of a minor quality degradation. **Disclaimer**: Content for this model card has partly been written by the 🤗 Hugging Face team, and partly copied and pasted from the original model card. ## Usage Whisper large-v3-turbo is supported in Hugging Face 🤗 Transformers. To run the model, first install the Transformers library. For this example, we'll also install 🤗 Datasets to load toy audio dataset from the Hugging Face Hub, and 🤗 Accelerate to reduce the model loading time: ```bash pip install --upgrade pip pip install --upgrade transformers datasets[audio] accelerate ``` The model can be used with the [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) class to transcribe audios of arbitrary length: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` To transcribe a local audio file, simply pass the path to your audio file when you call the pipeline: ```python result = pipe("audio.mp3") ``` Multiple audio files can be transcribed in parallel by specifying them as a list and setting the `batch_size` parameter: ```python result = pipe(["audio_1.mp3", "audio_2.mp3"], batch_size=2) ``` Transformers is compatible with all Whisper decoding strategies, such as temperature fallback and condition on previous tokens. The following example demonstrates how to enable these heuristics: ```python generate_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } result = pipe(sample, generate_kwargs=generate_kwargs) ``` Whisper predicts the language of the source audio automatically. If the source audio language is known *a-priori*, it can be passed as an argument to the pipeline: ```python result = pipe(sample, generate_kwargs={"language": "english"}) ``` By default, Whisper performs the task of *speech transcription*, where the source audio language is the same as the target text language. To perform *speech translation*, where the target text is in English, set the task to `"translate"`: ```python result = pipe(sample, generate_kwargs={"task": "translate"}) ``` Finally, the model can be made to predict timestamps. For sentence-level timestamps, pass the `return_timestamps` argument: ```python result = pipe(sample, return_timestamps=True) print(result["chunks"]) ``` And for word-level timestamps: ```python result = pipe(sample, return_timestamps="word") print(result["chunks"]) ``` The above arguments can be used in isolation or in combination. For example, to perform the task of speech transcription where the source audio is in French, and we want to return sentence-level timestamps, the following can be used: ```python result = pipe(sample, return_timestamps=True, generate_kwargs={"language": "french", "task": "translate"}) print(result["chunks"]) ``` <details> <summary> For more control over the generation parameters, use the model + processor API directly: </summary> ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor from datasets import Audio, load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") dataset = dataset.cast_column("audio", Audio(processor.feature_extractor.sampling_rate)) sample = dataset[0]["audio"] inputs = processor( sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt", truncation=False, padding="longest", return_attention_mask=True, ) inputs = inputs.to(device, dtype=torch_dtype) gen_kwargs = { "max_new_tokens": 448, "num_beams": 1, "condition_on_prev_tokens": False, "compression_ratio_threshold": 1.35, # zlib compression ratio threshold (in token space) "temperature": (0.0, 0.2, 0.4, 0.6, 0.8, 1.0), "logprob_threshold": -1.0, "no_speech_threshold": 0.6, "return_timestamps": True, } pred_ids = model.generate(**inputs, **gen_kwargs) pred_text = processor.batch_decode(pred_ids, skip_special_tokens=True, decode_with_timestamps=False) print(pred_text) ``` </details> ## Additional Speed & Memory Improvements You can apply additional speed and memory improvements to Whisper to further reduce the inference speed and VRAM requirements. ### Chunked Long-Form Whisper has a receptive field of 30-seconds. To transcribe audios longer than this, one of two long-form algorithms are required: 1. **Sequential:** uses a "sliding window" for buffered inference, transcribing 30-second slices one after the other 2. **Chunked:** splits long audio files into shorter ones (with a small overlap between segments), transcribes each segment independently, and stitches the resulting transcriptions at the boundaries The sequential long-form algorithm should be used in either of the following scenarios: 1. Transcription accuracy is the most important factor, and speed is less of a consideration 2. You are transcribing **batches** of long audio files, in which case the latency of sequential is comparable to chunked, while being up to 0.5% WER more accurate Conversely, the chunked algorithm should be used when: 1. Transcription speed is the most important factor 2. You are transcribing a **single** long audio file By default, Transformers uses the sequential algorithm. To enable the chunked algorithm, pass the `chunk_length_s` parameter to the `pipeline`. For large-v3, a chunk length of 30-seconds is optimal. To activate batching over long audio files, pass the argument `batch_size`: ```python import torch from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ) model.to(device) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, chunk_length_s=30, batch_size=16, # batch size for inference - set based on your device torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] result = pipe(sample) print(result["text"]) ``` #### Torch compile The Whisper forward pass is compatible with [`torch.compile`](https://pytorch.org/docs/stable/generated/torch.compile.html) for 4.5x speed-ups. **Note:** `torch.compile` is currently not compatible with the Chunked long-form algorithm or Flash Attention 2 ⚠️ ```python import torch from torch.nn.attention import SDPBackend, sdpa_kernel from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline from datasets import load_dataset from tqdm import tqdm torch.set_float32_matmul_precision("high") device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model_id = "ylacombe/whisper-large-v3-turbo" model = AutoModelForSpeechSeq2Seq.from_pretrained( model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True ).to(device) # Enable static cache and compile the forward pass model.generation_config.cache_implementation = "static" model.generation_config.max_new_tokens = 256 model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) processor = AutoProcessor.from_pretrained(model_id) pipe = pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, torch_dtype=torch_dtype, device=device, ) dataset = load_dataset("distil-whisper/librispeech_long", "clean", split="validation") sample = dataset[0]["audio"] # 2 warmup steps for _ in tqdm(range(2), desc="Warm-up step"): with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy(), generate_kwargs={"min_new_tokens": 256, "max_new_tokens": 256}) # fast run with sdpa_kernel(SDPBackend.MATH): result = pipe(sample.copy()) print(result["text"]) ``` #### Flash Attention 2 We recommend using [Flash-Attention 2](https://huggingface.co/docs/transformers/main/en/perf_infer_gpu_one#flashattention-2) if your GPU supports it and you are not using [torch.compile](#torch-compile). To do so, first install [Flash Attention](https://github.com/Dao-AILab/flash-attention): ``` pip install flash-attn --no-build-isolation ``` Then pass `attn_implementation="flash_attention_2"` to `from_pretrained`: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="flash_attention_2") ``` #### Torch Scale-Product-Attention (SDPA) If your GPU does not support Flash Attention, we recommend making use of PyTorch [scaled dot-product attention (SDPA)](https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html). This attention implementation is activated **by default** for PyTorch versions 2.1.1 or greater. To check whether you have a compatible PyTorch version, run the following Python code snippet: ```python from transformers.utils import is_torch_sdpa_available print(is_torch_sdpa_available()) ``` If the above returns `True`, you have a valid version of PyTorch installed and SDPA is activated by default. If it returns `False`, you need to upgrade your PyTorch version according to the [official instructions](https://pytorch.org/get-started/locally/) Once a valid PyTorch version is installed, SDPA is activated by default. It can also be set explicitly by specifying `attn_implementation="sdpa"` as follows: ```python model = AutoModelForSpeechSeq2Seq.from_pretrained(model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, attn_implementation="sdpa") ``` For more information about how to use the SDPA refer to the [Transformers SDPA documentation](https://huggingface.co/docs/transformers/en/perf_infer_gpu_one#pytorch-scaled-dot-product-attention). ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. There are two flavours of Whisper model: English-only and multilingual. The English-only models were trained on the task of English speech recognition. The multilingual models were trained simultaneously on multilingual speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the *same* language as the audio. For speech translation, the model predicts transcriptions to a *different* language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are available as English-only and multilingual. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: | Size | Parameters | English-only | Multilingual | |----------|------------|------------------------------------------------------|-----------------------------------------------------| | tiny | 39 M | [✓](https://huggingface.co/openai/whisper-tiny.en) | [✓](https://huggingface.co/openai/whisper-tiny) | | base | 74 M | [✓](https://huggingface.co/openai/whisper-base.en) | [✓](https://huggingface.co/openai/whisper-base) | | small | 244 M | [✓](https://huggingface.co/openai/whisper-small.en) | [✓](https://huggingface.co/openai/whisper-small) | | medium | 769 M | [✓](https://huggingface.co/openai/whisper-medium.en) | [✓](https://huggingface.co/openai/whisper-medium) | | large | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large) | | large-v2 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v2) | | large-v3 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v3) | | large-v3-turbo | 809 M | x | [✓](https://huggingface.co/ylacombe/whisper-large-v3-turbo) | ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through *fine-tuning*. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data No information provided. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```

{"language": ["en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", false, "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su"], "license": "apache-2.0", "pipeline_tag": "automatic-speech-recognition", "tags": ["audio", "automatic-speech-recognition", "hf-asr-leaderboard"], "widget": [{"example_title": "Librispeech sample 1", "src": "https://cdn-media.huggingface.co/speech_samples/sample1.flac"}, {"example_title": "Librispeech sample 2", "src": "https://cdn-media.huggingface.co/speech_samples/sample2.flac"}]}

task

[ "TRANSLATION" ]

44,535

Helsinki-NLP/opus-mt-tc-big-gmq-itc

Helsinki-NLP

translation

[ "transformers", "pytorch", "tf", "safetensors", "marian", "text2text-generation", "translation", "opus-mt-tc", "ca", "da", "es", "fr", "gl", "is", "it", "nb", "pt", "ro", "sv", "license:cc-by-4.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-08-12T11:15:27Z

2023-10-10T11:14:55+00:00

24

0

--- language: - ca - da - es - fr - gl - is - it - nb - pt - ro - sv license: cc-by-4.0 tags: - translation - opus-mt-tc model-index: - name: opus-mt-tc-big-gmq-itc results: - task: type: translation name: Translation dan-cat dataset: name: flores101-devtest type: flores_101 args: dan cat devtest metrics: - type: bleu value: 33.4 name: BLEU - type: chrf value: 0.59224 name: chr-F - type: bleu value: 38.3 name: BLEU - type: chrf value: 0.63387 name: chr-F - type: bleu value: 26.4 name: BLEU - type: chrf value: 0.54446 name: chr-F - type: bleu value: 25.7 name: BLEU - type: chrf value: 0.55237 name: chr-F - type: bleu value: 36.9 name: BLEU - type: chrf value: 0.62233 name: chr-F - type: bleu value: 31.8 name: BLEU - type: chrf value: 0.58235 name: chr-F - type: bleu value: 24.3 name: BLEU - type: chrf value: 0.52453 name: chr-F - type: bleu value: 22.7 name: BLEU - type: chrf value: 0.4893 name: chr-F - type: bleu value: 26.2 name: BLEU - type: chrf value: 0.52704 name: chr-F - type: bleu value: 18.0 name: BLEU - type: chrf value: 0.45387 name: chr-F - type: bleu value: 18.6 name: BLEU - type: chrf value: 0.47303 name: chr-F - type: bleu value: 24.9 name: BLEU - type: chrf value: 0.51381 name: chr-F - type: bleu value: 21.6 name: BLEU - type: chrf value: 0.48224 name: chr-F - type: bleu value: 18.1 name: BLEU - type: chrf value: 0.45786 name: chr-F - type: bleu value: 28.9 name: BLEU - type: chrf value: 0.55984 name: chr-F - type: bleu value: 33.8 name: BLEU - type: chrf value: 0.60102 name: chr-F - type: bleu value: 23.4 name: BLEU - type: chrf value: 0.52145 name: chr-F - type: bleu value: 22.2 name: BLEU - type: chrf value: 0.52619 name: chr-F - type: bleu value: 32.2 name: BLEU - type: chrf value: 0.58836 name: chr-F - type: bleu value: 27.6 name: BLEU - type: chrf value: 0.54845 name: chr-F - type: bleu value: 21.8 name: BLEU - type: chrf value: 0.50661 name: chr-F - type: bleu value: 32.4 name: BLEU - type: chrf value: 0.58542 name: chr-F - type: bleu value: 39.3 name: BLEU - type: chrf value: 0.63688 name: chr-F - type: bleu value: 26.0 name: BLEU - type: chrf value: 0.53989 name: chr-F - type: bleu value: 25.9 name: BLEU - type: chrf value: 0.55232 name: chr-F - type: bleu value: 36.5 name: BLEU - type: chrf value: 0.61882 name: chr-F - type: bleu value: 31.0 name: BLEU - type: chrf value: 0.57419 name: chr-F - type: bleu value: 23.8 name: BLEU - type: chrf value: 0.52175 name: chr-F - task: type: translation name: Translation dan-fra dataset: name: tatoeba-test-v2021-08-07 type: tatoeba_mt args: dan-fra metrics: - type: bleu value: 63.8 name: BLEU - type: chrf value: 0.76671 name: chr-F - type: bleu value: 56.2 name: BLEU - type: chrf value: 0.74658 name: chr-F - type: bleu value: 57.8 name: BLEU - type: chrf value: 0.74944 name: chr-F - type: bleu value: 54.8 name: BLEU - type: chrf value: 0.72328 name: chr-F - type: bleu value: 51.0 name: BLEU - type: chrf value: 0.69354 name: chr-F - type: bleu value: 49.2 name: BLEU - type: chrf value: 0.66008 name: chr-F - type: bleu value: 54.4 name: BLEU - type: chrf value: 0.70854 name: chr-F - type: bleu value: 55.9 name: BLEU - type: chrf value: 0.73672 name: chr-F - type: bleu value: 59.2 name: BLEU - type: chrf value: 0.73014 name: chr-F - type: bleu value: 56.6 name: BLEU - type: chrf value: 0.73211 name: chr-F - type: bleu value: 48.7 name: BLEU - type: chrf value: 0.68146 name: chr-F - type: bleu value: 55.3 name: BLEU - type: chrf value: 0.71373 name: chr-F --- # opus-mt-tc-big-gmq-itc ## Table of Contents - [Model Details](#model-details) - [Uses](#uses) - [Risks, Limitations and Biases](#risks-limitations-and-biases) - [How to Get Started With the Model](#how-to-get-started-with-the-model) - [Training](#training) - [Evaluation](#evaluation) - [Citation Information](#citation-information) - [Acknowledgements](#acknowledgements) ## Model Details Neural machine translation model for translating from North Germanic languages (gmq) to Italic languages (itc). This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train). **Model Description:** - **Developed by:** Language Technology Research Group at the University of Helsinki - **Model Type:** Translation (transformer-big) - **Release**: 2022-08-09 - **License:** CC-BY-4.0 - **Language(s):** - Source Language(s): dan isl nno nob nor swe - Target Language(s): cat fra glg ita lat por ron spa - Language Pair(s): dan-cat dan-fra dan-glg dan-ita dan-por dan-ron dan-spa isl-cat isl-fra isl-ita isl-por isl-ron isl-spa nob-cat nob-fra nob-glg nob-ita nob-por nob-ron nob-spa swe-cat swe-fra swe-glg swe-ita swe-por swe-ron swe-spa - Valid Target Language Labels: >>acf<< >>aoa<< >>arg<< >>ast<< >>cat<< >>cbk<< >>ccd<< >>cks<< >>cos<< >>cri<< >>crs<< >>dlm<< >>drc<< >>egl<< >>ext<< >>fab<< >>fax<< >>fra<< >>frc<< >>frm<< >>fro<< >>frp<< >>fur<< >>gcf<< >>gcr<< >>glg<< >>hat<< >>idb<< >>ist<< >>ita<< >>itk<< >>kea<< >>kmv<< >>lad<< >>lad_Latn<< >>lat<< >>lat_Latn<< >>lij<< >>lld<< >>lmo<< >>lou<< >>mcm<< >>mfe<< >>mol<< >>mwl<< >>mxi<< >>mzs<< >>nap<< >>nrf<< >>oci<< >>osc<< >>osp<< >>osp_Latn<< >>pap<< >>pcd<< >>pln<< >>pms<< >>pob<< >>por<< >>pov<< >>pre<< >>pro<< >>qbb<< >>qhr<< >>rcf<< >>rgn<< >>roh<< >>ron<< >>ruo<< >>rup<< >>ruq<< >>scf<< >>scn<< >>sdc<< >>sdn<< >>spa<< >>spq<< >>spx<< >>src<< >>srd<< >>sro<< >>tmg<< >>tvy<< >>vec<< >>vkp<< >>wln<< >>xfa<< >>xum<< - **Original Model**: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - **Resources for more information:** - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) - More information about released models for this language pair: [OPUS-MT gmq-itc README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/gmq-itc/README.md) - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian) - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/ This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>fra<<` ## Uses This model can be used for translation and text-to-text generation. ## Risks, Limitations and Biases **CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes.** Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). ## How to Get Started With the Model A short example code: ```python from transformers import MarianMTModel, MarianTokenizer src_text = [ ">>spa<< Jag är inte religiös.", ">>por<< Livet er for kort til å lære seg tysk." ] model_name = "pytorch-models/opus-mt-tc-big-gmq-itc" tokenizer = MarianTokenizer.from_pretrained(model_name) model = MarianMTModel.from_pretrained(model_name) translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True)) for t in translated: print( tokenizer.decode(t, skip_special_tokens=True) ) # expected output: # No soy religioso. # A vida é muito curta para aprender alemão. ``` You can also use OPUS-MT models with the transformers pipelines, for example: ```python from transformers import pipeline pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-gmq-itc") print(pipe(">>spa<< Jag är inte religiös.")) # expected output: No soy religioso. ``` ## Training - **Data**: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge)) - **Pre-processing**: SentencePiece (spm32k,spm32k) - **Model Type:** transformer-big - **Original MarianNMT Model**: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) ## Evaluation * test set translations: [opusTCv20210807_transformer-big_2022-08-09.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.test.txt) * test set scores: [opusTCv20210807_transformer-big_2022-08-09.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.eval.txt) * benchmark results: [benchmark_results.txt](benchmark_results.txt) * benchmark output: [benchmark_translations.zip](benchmark_translations.zip) | langpair | testset | chr-F | BLEU | #sent | #words | |----------|---------|-------|-------|-------|--------| | dan-fra | tatoeba-test-v2021-08-07 | 0.76671 | 63.8 | 1731 | 11882 | | dan-ita | tatoeba-test-v2021-08-07 | 0.74658 | 56.2 | 284 | 2226 | | dan-por | tatoeba-test-v2021-08-07 | 0.74944 | 57.8 | 873 | 5360 | | dan-spa | tatoeba-test-v2021-08-07 | 0.72328 | 54.8 | 5000 | 35528 | | isl-ita | tatoeba-test-v2021-08-07 | 0.69354 | 51.0 | 236 | 1450 | | isl-spa | tatoeba-test-v2021-08-07 | 0.66008 | 49.2 | 238 | 1229 | | nob-fra | tatoeba-test-v2021-08-07 | 0.70854 | 54.4 | 323 | 2269 | | nob-spa | tatoeba-test-v2021-08-07 | 0.73672 | 55.9 | 885 | 6866 | | swe-fra | tatoeba-test-v2021-08-07 | 0.73014 | 59.2 | 1407 | 9580 | | swe-ita | tatoeba-test-v2021-08-07 | 0.73211 | 56.6 | 715 | 4711 | | swe-por | tatoeba-test-v2021-08-07 | 0.68146 | 48.7 | 320 | 2032 | | swe-spa | tatoeba-test-v2021-08-07 | 0.71373 | 55.3 | 1351 | 8235 | | dan-cat | flores101-devtest | 0.59224 | 33.4 | 1012 | 27304 | | dan-fra | flores101-devtest | 0.63387 | 38.3 | 1012 | 28343 | | dan-glg | flores101-devtest | 0.54446 | 26.4 | 1012 | 26582 | | dan-ita | flores101-devtest | 0.55237 | 25.7 | 1012 | 27306 | | dan-por | flores101-devtest | 0.62233 | 36.9 | 1012 | 26519 | | dan-ron | flores101-devtest | 0.58235 | 31.8 | 1012 | 26799 | | dan-spa | flores101-devtest | 0.52453 | 24.3 | 1012 | 29199 | | isl-cat | flores101-devtest | 0.48930 | 22.7 | 1012 | 27304 | | isl-fra | flores101-devtest | 0.52704 | 26.2 | 1012 | 28343 | | isl-glg | flores101-devtest | 0.45387 | 18.0 | 1012 | 26582 | | isl-ita | flores101-devtest | 0.47303 | 18.6 | 1012 | 27306 | | isl-por | flores101-devtest | 0.51381 | 24.9 | 1012 | 26519 | | isl-ron | flores101-devtest | 0.48224 | 21.6 | 1012 | 26799 | | isl-spa | flores101-devtest | 0.45786 | 18.1 | 1012 | 29199 | | nob-cat | flores101-devtest | 0.55984 | 28.9 | 1012 | 27304 | | nob-fra | flores101-devtest | 0.60102 | 33.8 | 1012 | 28343 | | nob-glg | flores101-devtest | 0.52145 | 23.4 | 1012 | 26582 | | nob-ita | flores101-devtest | 0.52619 | 22.2 | 1012 | 27306 | | nob-por | flores101-devtest | 0.58836 | 32.2 | 1012 | 26519 | | nob-ron | flores101-devtest | 0.54845 | 27.6 | 1012 | 26799 | | nob-spa | flores101-devtest | 0.50661 | 21.8 | 1012 | 29199 | | swe-cat | flores101-devtest | 0.58542 | 32.4 | 1012 | 27304 | | swe-fra | flores101-devtest | 0.63688 | 39.3 | 1012 | 28343 | | swe-glg | flores101-devtest | 0.53989 | 26.0 | 1012 | 26582 | | swe-ita | flores101-devtest | 0.55232 | 25.9 | 1012 | 27306 | | swe-por | flores101-devtest | 0.61882 | 36.5 | 1012 | 26519 | | swe-ron | flores101-devtest | 0.57419 | 31.0 | 1012 | 26799 | | swe-spa | flores101-devtest | 0.52175 | 23.8 | 1012 | 29199 | ## Citation Information * Publications: [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.) ``` @inproceedings{tiedemann-thottingal-2020-opus, title = "{OPUS}-{MT} {--} Building open translation services for the World", author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh}, booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation", month = nov, year = "2020", address = "Lisboa, Portugal", publisher = "European Association for Machine Translation", url = "https://aclanthology.org/2020.eamt-1.61", pages = "479--480", } @inproceedings{tiedemann-2020-tatoeba, title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}", author = {Tiedemann, J{\"o}rg}, booktitle = "Proceedings of the Fifth Conference on Machine Translation", month = nov, year = "2020", address = "Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2020.wmt-1.139", pages = "1174--1182", } ``` ## Acknowledgements The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland. ## Model conversion info * transformers version: 4.16.2 * OPUS-MT git hash: 8b9f0b0 * port time: Sat Aug 13 00:00:00 EEST 2022 * port machine: LM0-400-22516.local

null

Non_BioNLP

# opus-mt-tc-big-gmq-itc ## Table of Contents - [Model Details](#model-details) - [Uses](#uses) - [Risks, Limitations and Biases](#risks-limitations-and-biases) - [How to Get Started With the Model](#how-to-get-started-with-the-model) - [Training](#training) - [Evaluation](#evaluation) - [Citation Information](#citation-information) - [Acknowledgements](#acknowledgements) ## Model Details Neural machine translation model for translating from North Germanic languages (gmq) to Italic languages (itc). This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train). **Model Description:** - **Developed by:** Language Technology Research Group at the University of Helsinki - **Model Type:** Translation (transformer-big) - **Release**: 2022-08-09 - **License:** CC-BY-4.0 - **Language(s):** - Source Language(s): dan isl nno nob nor swe - Target Language(s): cat fra glg ita lat por ron spa - Language Pair(s): dan-cat dan-fra dan-glg dan-ita dan-por dan-ron dan-spa isl-cat isl-fra isl-ita isl-por isl-ron isl-spa nob-cat nob-fra nob-glg nob-ita nob-por nob-ron nob-spa swe-cat swe-fra swe-glg swe-ita swe-por swe-ron swe-spa - Valid Target Language Labels: >>acf<< >>aoa<< >>arg<< >>ast<< >>cat<< >>cbk<< >>ccd<< >>cks<< >>cos<< >>cri<< >>crs<< >>dlm<< >>drc<< >>egl<< >>ext<< >>fab<< >>fax<< >>fra<< >>frc<< >>frm<< >>fro<< >>frp<< >>fur<< >>gcf<< >>gcr<< >>glg<< >>hat<< >>idb<< >>ist<< >>ita<< >>itk<< >>kea<< >>kmv<< >>lad<< >>lad_Latn<< >>lat<< >>lat_Latn<< >>lij<< >>lld<< >>lmo<< >>lou<< >>mcm<< >>mfe<< >>mol<< >>mwl<< >>mxi<< >>mzs<< >>nap<< >>nrf<< >>oci<< >>osc<< >>osp<< >>osp_Latn<< >>pap<< >>pcd<< >>pln<< >>pms<< >>pob<< >>por<< >>pov<< >>pre<< >>pro<< >>qbb<< >>qhr<< >>rcf<< >>rgn<< >>roh<< >>ron<< >>ruo<< >>rup<< >>ruq<< >>scf<< >>scn<< >>sdc<< >>sdn<< >>spa<< >>spq<< >>spx<< >>src<< >>srd<< >>sro<< >>tmg<< >>tvy<< >>vec<< >>vkp<< >>wln<< >>xfa<< >>xum<< - **Original Model**: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - **Resources for more information:** - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) - More information about released models for this language pair: [OPUS-MT gmq-itc README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/gmq-itc/README.md) - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian) - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/ This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>fra<<` ## Uses This model can be used for translation and text-to-text generation. ## Risks, Limitations and Biases **CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes.** Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). ## How to Get Started With the Model A short example code: ```python from transformers import MarianMTModel, MarianTokenizer src_text = [ ">>spa<< Jag är inte religiös.", ">>por<< Livet er for kort til å lære seg tysk." ] model_name = "pytorch-models/opus-mt-tc-big-gmq-itc" tokenizer = MarianTokenizer.from_pretrained(model_name) model = MarianMTModel.from_pretrained(model_name) translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True)) for t in translated: print( tokenizer.decode(t, skip_special_tokens=True) ) # expected output: # No soy religioso. # A vida é muito curta para aprender alemão. ``` You can also use OPUS-MT models with the transformers pipelines, for example: ```python from transformers import pipeline pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-gmq-itc") print(pipe(">>spa<< Jag är inte religiös.")) # expected output: No soy religioso. ``` ## Training - **Data**: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge)) - **Pre-processing**: SentencePiece (spm32k,spm32k) - **Model Type:** transformer-big - **Original MarianNMT Model**: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip) - **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train) ## Evaluation * test set translations: [opusTCv20210807_transformer-big_2022-08-09.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.test.txt) * test set scores: [opusTCv20210807_transformer-big_2022-08-09.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.eval.txt) * benchmark results: [benchmark_results.txt](benchmark_results.txt) * benchmark output: [benchmark_translations.zip](benchmark_translations.zip) | langpair | testset | chr-F | BLEU | #sent | #words | |----------|---------|-------|-------|-------|--------| | dan-fra | tatoeba-test-v2021-08-07 | 0.76671 | 63.8 | 1731 | 11882 | | dan-ita | tatoeba-test-v2021-08-07 | 0.74658 | 56.2 | 284 | 2226 | | dan-por | tatoeba-test-v2021-08-07 | 0.74944 | 57.8 | 873 | 5360 | | dan-spa | tatoeba-test-v2021-08-07 | 0.72328 | 54.8 | 5000 | 35528 | | isl-ita | tatoeba-test-v2021-08-07 | 0.69354 | 51.0 | 236 | 1450 | | isl-spa | tatoeba-test-v2021-08-07 | 0.66008 | 49.2 | 238 | 1229 | | nob-fra | tatoeba-test-v2021-08-07 | 0.70854 | 54.4 | 323 | 2269 | | nob-spa | tatoeba-test-v2021-08-07 | 0.73672 | 55.9 | 885 | 6866 | | swe-fra | tatoeba-test-v2021-08-07 | 0.73014 | 59.2 | 1407 | 9580 | | swe-ita | tatoeba-test-v2021-08-07 | 0.73211 | 56.6 | 715 | 4711 | | swe-por | tatoeba-test-v2021-08-07 | 0.68146 | 48.7 | 320 | 2032 | | swe-spa | tatoeba-test-v2021-08-07 | 0.71373 | 55.3 | 1351 | 8235 | | dan-cat | flores101-devtest | 0.59224 | 33.4 | 1012 | 27304 | | dan-fra | flores101-devtest | 0.63387 | 38.3 | 1012 | 28343 | | dan-glg | flores101-devtest | 0.54446 | 26.4 | 1012 | 26582 | | dan-ita | flores101-devtest | 0.55237 | 25.7 | 1012 | 27306 | | dan-por | flores101-devtest | 0.62233 | 36.9 | 1012 | 26519 | | dan-ron | flores101-devtest | 0.58235 | 31.8 | 1012 | 26799 | | dan-spa | flores101-devtest | 0.52453 | 24.3 | 1012 | 29199 | | isl-cat | flores101-devtest | 0.48930 | 22.7 | 1012 | 27304 | | isl-fra | flores101-devtest | 0.52704 | 26.2 | 1012 | 28343 | | isl-glg | flores101-devtest | 0.45387 | 18.0 | 1012 | 26582 | | isl-ita | flores101-devtest | 0.47303 | 18.6 | 1012 | 27306 | | isl-por | flores101-devtest | 0.51381 | 24.9 | 1012 | 26519 | | isl-ron | flores101-devtest | 0.48224 | 21.6 | 1012 | 26799 | | isl-spa | flores101-devtest | 0.45786 | 18.1 | 1012 | 29199 | | nob-cat | flores101-devtest | 0.55984 | 28.9 | 1012 | 27304 | | nob-fra | flores101-devtest | 0.60102 | 33.8 | 1012 | 28343 | | nob-glg | flores101-devtest | 0.52145 | 23.4 | 1012 | 26582 | | nob-ita | flores101-devtest | 0.52619 | 22.2 | 1012 | 27306 | | nob-por | flores101-devtest | 0.58836 | 32.2 | 1012 | 26519 | | nob-ron | flores101-devtest | 0.54845 | 27.6 | 1012 | 26799 | | nob-spa | flores101-devtest | 0.50661 | 21.8 | 1012 | 29199 | | swe-cat | flores101-devtest | 0.58542 | 32.4 | 1012 | 27304 | | swe-fra | flores101-devtest | 0.63688 | 39.3 | 1012 | 28343 | | swe-glg | flores101-devtest | 0.53989 | 26.0 | 1012 | 26582 | | swe-ita | flores101-devtest | 0.55232 | 25.9 | 1012 | 27306 | | swe-por | flores101-devtest | 0.61882 | 36.5 | 1012 | 26519 | | swe-ron | flores101-devtest | 0.57419 | 31.0 | 1012 | 26799 | | swe-spa | flores101-devtest | 0.52175 | 23.8 | 1012 | 29199 | ## Citation Information * Publications: [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.) ``` @inproceedings{tiedemann-thottingal-2020-opus, title = "{OPUS}-{MT} {--} Building open translation services for the World", author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh}, booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation", month = nov, year = "2020", address = "Lisboa, Portugal", publisher = "European Association for Machine Translation", url = "https://aclanthology.org/2020.eamt-1.61", pages = "479--480", } @inproceedings{tiedemann-2020-tatoeba, title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}", author = {Tiedemann, J{\"o}rg}, booktitle = "Proceedings of the Fifth Conference on Machine Translation", month = nov, year = "2020", address = "Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2020.wmt-1.139", pages = "1174--1182", } ``` ## Acknowledgements The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland. ## Model conversion info * transformers version: 4.16.2 * OPUS-MT git hash: 8b9f0b0 * port time: Sat Aug 13 00:00:00 EEST 2022 * port machine: LM0-400-22516.local

{"language": ["ca", "da", "es", "fr", "gl", "is", "it", "nb", "pt", "ro", "sv"], "license": "cc-by-4.0", "tags": ["translation", "opus-mt-tc"], "model-index": [{"name": "opus-mt-tc-big-gmq-itc", "results": [{"task": {"type": "translation", "name": "Translation dan-cat"}, "dataset": {"name": "flores101-devtest", "type": "flores_101", "args": "dan cat devtest"}, "metrics": [{"type": "bleu", "value": 33.4, "name": "BLEU"}, {"type": "chrf", "value": 0.59224, "name": "chr-F"}, {"type": "bleu", "value": 38.3, "name": "BLEU"}, {"type": "chrf", "value": 0.63387, "name": "chr-F"}, {"type": "bleu", "value": 26.4, "name": "BLEU"}, {"type": "chrf", "value": 0.54446, "name": "chr-F"}, {"type": "bleu", "value": 25.7, "name": "BLEU"}, {"type": "chrf", "value": 0.55237, "name": "chr-F"}, {"type": "bleu", "value": 36.9, "name": "BLEU"}, {"type": "chrf", "value": 0.62233, "name": "chr-F"}, {"type": "bleu", "value": 31.8, "name": "BLEU"}, {"type": "chrf", "value": 0.58235, "name": "chr-F"}, {"type": "bleu", "value": 24.3, "name": "BLEU"}, {"type": "chrf", "value": 0.52453, "name": "chr-F"}, {"type": "bleu", "value": 22.7, "name": "BLEU"}, {"type": "chrf", "value": 0.4893, "name": "chr-F"}, {"type": "bleu", "value": 26.2, "name": "BLEU"}, {"type": "chrf", "value": 0.52704, "name": "chr-F"}, {"type": "bleu", "value": 18.0, "name": "BLEU"}, {"type": "chrf", "value": 0.45387, "name": "chr-F"}, {"type": "bleu", "value": 18.6, "name": "BLEU"}, {"type": "chrf", "value": 0.47303, "name": "chr-F"}, {"type": "bleu", "value": 24.9, "name": "BLEU"}, {"type": "chrf", "value": 0.51381, "name": "chr-F"}, {"type": "bleu", "value": 21.6, "name": "BLEU"}, {"type": "chrf", "value": 0.48224, "name": "chr-F"}, {"type": "bleu", "value": 18.1, "name": "BLEU"}, {"type": "chrf", "value": 0.45786, "name": "chr-F"}, {"type": "bleu", "value": 28.9, "name": "BLEU"}, {"type": "chrf", "value": 0.55984, "name": "chr-F"}, {"type": "bleu", "value": 33.8, "name": "BLEU"}, {"type": "chrf", "value": 0.60102, "name": "chr-F"}, {"type": "bleu", "value": 23.4, "name": "BLEU"}, {"type": "chrf", "value": 0.52145, "name": "chr-F"}, {"type": "bleu", "value": 22.2, "name": "BLEU"}, {"type": "chrf", "value": 0.52619, "name": "chr-F"}, {"type": "bleu", "value": 32.2, "name": "BLEU"}, {"type": "chrf", "value": 0.58836, "name": "chr-F"}, {"type": "bleu", "value": 27.6, "name": "BLEU"}, {"type": "chrf", "value": 0.54845, "name": "chr-F"}, {"type": "bleu", "value": 21.8, "name": "BLEU"}, {"type": "chrf", "value": 0.50661, "name": "chr-F"}, {"type": "bleu", "value": 32.4, "name": "BLEU"}, {"type": "chrf", "value": 0.58542, "name": "chr-F"}, {"type": "bleu", "value": 39.3, "name": "BLEU"}, {"type": "chrf", "value": 0.63688, "name": "chr-F"}, {"type": "bleu", "value": 26.0, "name": "BLEU"}, {"type": "chrf", "value": 0.53989, "name": "chr-F"}, {"type": "bleu", "value": 25.9, "name": "BLEU"}, {"type": "chrf", "value": 0.55232, "name": "chr-F"}, {"type": "bleu", "value": 36.5, "name": "BLEU"}, {"type": "chrf", "value": 0.61882, "name": "chr-F"}, {"type": "bleu", "value": 31.0, "name": "BLEU"}, {"type": "chrf", "value": 0.57419, "name": "chr-F"}, {"type": "bleu", "value": 23.8, "name": "BLEU"}, {"type": "chrf", "value": 0.52175, "name": "chr-F"}]}, {"task": {"type": "translation", "name": "Translation dan-fra"}, "dataset": {"name": "tatoeba-test-v2021-08-07", "type": "tatoeba_mt", "args": "dan-fra"}, "metrics": [{"type": "bleu", "value": 63.8, "name": "BLEU"}, {"type": "chrf", "value": 0.76671, "name": "chr-F"}, {"type": "bleu", "value": 56.2, "name": "BLEU"}, {"type": "chrf", "value": 0.74658, "name": "chr-F"}, {"type": "bleu", "value": 57.8, "name": "BLEU"}, {"type": "chrf", "value": 0.74944, "name": "chr-F"}, {"type": "bleu", "value": 54.8, "name": "BLEU"}, {"type": "chrf", "value": 0.72328, "name": "chr-F"}, {"type": "bleu", "value": 51.0, "name": "BLEU"}, {"type": "chrf", "value": 0.69354, "name": "chr-F"}, {"type": "bleu", "value": 49.2, "name": "BLEU"}, {"type": "chrf", "value": 0.66008, "name": "chr-F"}, {"type": "bleu", "value": 54.4, "name": "BLEU"}, {"type": "chrf", "value": 0.70854, "name": "chr-F"}, {"type": "bleu", "value": 55.9, "name": "BLEU"}, {"type": "chrf", "value": 0.73672, "name": "chr-F"}, {"type": "bleu", "value": 59.2, "name": "BLEU"}, {"type": "chrf", "value": 0.73014, "name": "chr-F"}, {"type": "bleu", "value": 56.6, "name": "BLEU"}, {"type": "chrf", "value": 0.73211, "name": "chr-F"}, {"type": "bleu", "value": 48.7, "name": "BLEU"}, {"type": "chrf", "value": 0.68146, "name": "chr-F"}, {"type": "bleu", "value": 55.3, "name": "BLEU"}, {"type": "chrf", "value": 0.71373, "name": "chr-F"}]}]}]}

task

[ "TRANSLATION" ]

44,536

phospho-app/phospho-small-c99c7a6

phospho-app

feature-extraction

[ "transformers", "safetensors", "mpnet", "feature-extraction", "en", "license:apache-2.0", "endpoints_compatible", "region:us" ]

2024-04-30T08:38:30Z

2024-04-30T08:39:49+00:00

7

0

--- language: en license: apache-2.0 --- # phospho-small This is a SetFit model that can be used for Text Classification on CPU. The model has been trained using an efficient few-shot learning technique. ## Usage ```python from setfit import SetFitModel model = SetFitModel.from_pretrained("phospho-small-c99c7a6") outputs = model.predict(["This is a sentence to classify", "Another sentence"]) # tensor([1, 0]) ``` ## References This work was possible thanks to the SetFit library and the work of: Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren (2022). Efficient Few-Shot Learning Without Prompts. ArXiv: [https://doi.org/10.48550/arxiv.2209.11055](https://doi.org/10.48550/arxiv.2209.11055)

null

Non_BioNLP

# phospho-small This is a SetFit model that can be used for Text Classification on CPU. The model has been trained using an efficient few-shot learning technique. ## Usage ```python from setfit import SetFitModel model = SetFitModel.from_pretrained("phospho-small-c99c7a6") outputs = model.predict(["This is a sentence to classify", "Another sentence"]) # tensor([1, 0]) ``` ## References This work was possible thanks to the SetFit library and the work of: Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren (2022). Efficient Few-Shot Learning Without Prompts. ArXiv: [https://doi.org/10.48550/arxiv.2209.11055](https://doi.org/10.48550/arxiv.2209.11055)

{"language": "en", "license": "apache-2.0"}

task

[ "TEXT_CLASSIFICATION" ]

44,537

tosin/pcl_22

tosin

text2text-generation

[ "transformers", "pytorch", "t5", "text2text-generation", "text classification", "en", "dataset:PCL", "license:cc-by-4.0", "autotrain_compatible", "text-generation-inference", "region:us" ]

2022-03-02T23:29:05Z

2022-02-18T12:33:52+00:00

174

0

--- datasets: - PCL language: - en license: cc-by-4.0 metrics: - F1 tags: - text classification - transformers thumbnail: https://huggingface.co/front/thumbnails/dialogpt.png inference: false --- ## T5Base-PCL This is a fine-tuned model of T5 (base) on the patronizing and condenscending language (PCL) dataset by Pérez-Almendros et al (2020) used for Task 4 competition of SemEval-2022. It is intended to be used as a classification model for identifying PCL (0 - neg; 1 - pos). The task prefix we used for the T5 model is 'classification: '. The dataset it's trained on is limited in scope, as it covers only some news texts covering about 20 English-speaking countries. The macro F1 score achieved on the test set, based on the official evaluation, is 0.5452. More information about the original pre-trained model can be found [here](https://huggingface.co/t5-base) * Classification examples: |Prediction | Input | |---------|------------| |0 | selective kindness : in europe , some refugees are more equal than others | |1 | he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty | ### How to use ```python from transformers import T5ForConditionalGeneration, T5Tokenizer import torch model = T5ForConditionalGeneration.from_pretrained("tosin/pcl_22") tokenizer = T5Tokenizer.from_pretrained("t5-base") # use the source tokenizer because T5 finetuned tokenizer breaks tokenizer.pad_token = tokenizer.eos_token input_ids = tokenizer("he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty", padding=True, truncation=True, return_tensors='pt').input_ids outputs = model.generate(input_ids) pred = tokenizer.decode(outputs[0], skip_special_tokens=True) print(pred)

null

TBD

## T5Base-PCL This is a fine-tuned model of T5 (base) on the patronizing and condenscending language (PCL) dataset by Pérez-Almendros et al (2020) used for Task 4 competition of SemEval-2022. It is intended to be used as a classification model for identifying PCL (0 - neg; 1 - pos). The task prefix we used for the T5 model is 'classification: '. The dataset it's trained on is limited in scope, as it covers only some news texts covering about 20 English-speaking countries. The macro F1 score achieved on the test set, based on the official evaluation, is 0.5452. More information about the original pre-trained model can be found [here](https://huggingface.co/t5-base) * Classification examples: |Prediction | Input | |---------|------------| |0 | selective kindness : in europe , some refugees are more equal than others | |1 | he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty | ### How to use ```python from transformers import T5ForConditionalGeneration, T5Tokenizer import torch model = T5ForConditionalGeneration.from_pretrained("tosin/pcl_22") tokenizer = T5Tokenizer.from_pretrained("t5-base") # use the source tokenizer because T5 finetuned tokenizer breaks tokenizer.pad_token = tokenizer.eos_token input_ids = tokenizer("he said their efforts should not stop only at creating many graduates but also extended to students from poor families so that they could break away from the cycle of poverty", padding=True, truncation=True, return_tensors='pt').input_ids outputs = model.generate(input_ids) pred = tokenizer.decode(outputs[0], skip_special_tokens=True) print(pred)

{"datasets": ["PCL"], "language": ["en"], "license": "cc-by-4.0", "metrics": ["F1"], "tags": ["text classification", "transformers"], "thumbnail": "https://huggingface.co/front/thumbnails/dialogpt.png", "inference": false}

task

[ "TEXT_CLASSIFICATION" ]

44,538

gokuls/distilbert_sa_GLUE_Experiment_logit_kd_qnli

gokuls

text-classification

[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "en", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-01-29T22:09:31Z

2023-01-29T22:42:55+00:00

136

0

--- datasets: - glue language: - en license: apache-2.0 metrics: - accuracy tags: - generated_from_trainer model-index: - name: distilbert_sa_GLUE_Experiment_logit_kd_qnli results: - task: type: text-classification name: Text Classification dataset: name: GLUE QNLI type: glue config: qnli split: validation args: qnli metrics: - type: accuracy value: 0.5905180303862346 name: Accuracy ---  # distilbert_sa_GLUE_Experiment_logit_kd_qnli This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the GLUE QNLI dataset. It achieves the following results on the evaluation set: - Loss: 0.3865 - Accuracy: 0.5905 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - distributed_type: multi-GPU - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 50 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.4069 | 1.0 | 410 | 0.3914 | 0.5680 | | 0.3877 | 2.0 | 820 | 0.3865 | 0.5905 | | 0.3741 | 3.0 | 1230 | 0.3917 | 0.5971 | | 0.3604 | 4.0 | 1640 | 0.3893 | 0.5929 | | 0.3432 | 5.0 | 2050 | 0.3908 | 0.5922 | | 0.3194 | 6.0 | 2460 | 0.4251 | 0.5861 | | 0.2938 | 7.0 | 2870 | 0.4414 | 0.5940 | ### Framework versions - Transformers 4.26.0 - Pytorch 1.14.0a0+410ce96 - Datasets 2.9.0 - Tokenizers 0.13.2

null

Non_BioNLP

# distilbert_sa_GLUE_Experiment_logit_kd_qnli This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the GLUE QNLI dataset. It achieves the following results on the evaluation set: - Loss: 0.3865 - Accuracy: 0.5905 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - distributed_type: multi-GPU - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 50 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.4069 | 1.0 | 410 | 0.3914 | 0.5680 | | 0.3877 | 2.0 | 820 | 0.3865 | 0.5905 | | 0.3741 | 3.0 | 1230 | 0.3917 | 0.5971 | | 0.3604 | 4.0 | 1640 | 0.3893 | 0.5929 | | 0.3432 | 5.0 | 2050 | 0.3908 | 0.5922 | | 0.3194 | 6.0 | 2460 | 0.4251 | 0.5861 | | 0.2938 | 7.0 | 2870 | 0.4414 | 0.5940 | ### Framework versions - Transformers 4.26.0 - Pytorch 1.14.0a0+410ce96 - Datasets 2.9.0 - Tokenizers 0.13.2

{"datasets": ["glue"], "language": ["en"], "license": "apache-2.0", "metrics": ["accuracy"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert_sa_GLUE_Experiment_logit_kd_qnli", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "GLUE QNLI", "type": "glue", "config": "qnli", "split": "validation", "args": "qnli"}, "metrics": [{"type": "accuracy", "value": 0.5905180303862346, "name": "Accuracy"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,539

Alphatao/f2b0fa58-8c68-4ee4-8ff5-65c996bc267d

Alphatao

null

[ "peft", "safetensors", "llama", "axolotl", "generated_from_trainer", "base_model:llamafactory/tiny-random-Llama-3", "base_model:adapter:llamafactory/tiny-random-Llama-3", "license:apache-2.0", "region:us" ]

2025-03-10T07:37:37Z

2025-03-10T08:02:44+00:00

8

0

--- base_model: llamafactory/tiny-random-Llama-3 library_name: peft license: apache-2.0 tags: - axolotl - generated_from_trainer model-index: - name: f2b0fa58-8c68-4ee4-8ff5-65c996bc267d results: [] ---  [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: llamafactory/tiny-random-Llama-3 bf16: true chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 420d808b1b64ffc0_train_data.json ds_type: json format: custom path: /workspace/input_data/420d808b1b64ffc0_train_data.json type: field_instruction: source field_output: good-translation format: '{instruction}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null device_map: ? '' : 0,1,2,3,4,5,6,7 early_stopping_patience: 2 eval_max_new_tokens: 128 eval_steps: 100 eval_table_size: null flash_attention: true gradient_accumulation_steps: 8 gradient_checkpointing: true group_by_length: false hub_model_id: Alphatao/f2b0fa58-8c68-4ee4-8ff5-65c996bc267d hub_repo: null hub_strategy: null hub_token: null learning_rate: 0.0002 load_best_model_at_end: true load_in_4bit: false load_in_8bit: false local_rank: null logging_steps: 1 lora_alpha: 128 lora_dropout: 0.1 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 64 lora_target_linear: true lora_target_modules: - q_proj - k_proj - v_proj - o_proj lr_scheduler: cosine max_grad_norm: 1.0 max_steps: 5376 micro_batch_size: 4 mlflow_experiment_name: /tmp/420d808b1b64ffc0_train_data.json model_type: AutoModelForCausalLM num_epochs: 2 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false save_steps: 100 sequence_len: 2048 special_tokens: pad_token: <|eot_id|> strict: false tf32: true tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.04 wandb_entity: null wandb_mode: online wandb_name: 70c61e5c-5f68-4639-804d-89212f8c7742 wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 70c61e5c-5f68-4639-804d-89212f8c7742 warmup_steps: 10 weight_decay: 0.0 xformers_attention: null ``` </details> # f2b0fa58-8c68-4ee4-8ff5-65c996bc267d This model is a fine-tuned version of [llamafactory/tiny-random-Llama-3](https://huggingface.co/llamafactory/tiny-random-Llama-3) on the None dataset. It achieves the following results on the evaluation set: - Loss: 11.7233 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 32 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 10 - training_steps: 2123 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | 11.7614 | 0.0009 | 1 | 11.7621 | | 11.7349 | 0.0942 | 100 | 11.7364 | | 11.7389 | 0.1885 | 200 | 11.7346 | | 11.732 | 0.2827 | 300 | 11.7321 | | 11.7438 | 0.3769 | 400 | 11.7298 | | 11.7335 | 0.4711 | 500 | 11.7285 | | 11.7324 | 0.5654 | 600 | 11.7275 | | 11.7331 | 0.6596 | 700 | 11.7265 | | 11.7274 | 0.7538 | 800 | 11.7258 | | 11.7295 | 0.8481 | 900 | 11.7255 | | 11.731 | 0.9423 | 1000 | 11.7251 | | 13.0224 | 1.0365 | 1100 | 11.7249 | | 11.6643 | 1.1307 | 1200 | 11.7242 | | 11.7749 | 1.2250 | 1300 | 11.7240 | | 11.8175 | 1.3192 | 1400 | 11.7238 | | 10.3085 | 1.4134 | 1500 | 11.7236 | | 12.9048 | 1.5077 | 1600 | 11.7235 | | 10.6263 | 1.6019 | 1700 | 11.7234 | | 10.6834 | 1.6961 | 1800 | 11.7233 | | 11.5317 | 1.7903 | 1900 | 11.7233 | | 12.3242 | 1.8846 | 2000 | 11.7233 | | 10.9003 | 1.9788 | 2100 | 11.7233 | ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1

null

Non_BioNLP

[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml adapter: lora base_model: llamafactory/tiny-random-Llama-3 bf16: true chat_template: llama3 dataset_prepared_path: null datasets: - data_files: - 420d808b1b64ffc0_train_data.json ds_type: json format: custom path: /workspace/input_data/420d808b1b64ffc0_train_data.json type: field_instruction: source field_output: good-translation format: '{instruction}' no_input_format: '{instruction}' system_format: '{system}' system_prompt: '' debug: null device_map: ? '' : 0,1,2,3,4,5,6,7 early_stopping_patience: 2 eval_max_new_tokens: 128 eval_steps: 100 eval_table_size: null flash_attention: true gradient_accumulation_steps: 8 gradient_checkpointing: true group_by_length: false hub_model_id: Alphatao/f2b0fa58-8c68-4ee4-8ff5-65c996bc267d hub_repo: null hub_strategy: null hub_token: null learning_rate: 0.0002 load_best_model_at_end: true load_in_4bit: false load_in_8bit: false local_rank: null logging_steps: 1 lora_alpha: 128 lora_dropout: 0.1 lora_fan_in_fan_out: null lora_model_dir: null lora_r: 64 lora_target_linear: true lora_target_modules: - q_proj - k_proj - v_proj - o_proj lr_scheduler: cosine max_grad_norm: 1.0 max_steps: 5376 micro_batch_size: 4 mlflow_experiment_name: /tmp/420d808b1b64ffc0_train_data.json model_type: AutoModelForCausalLM num_epochs: 2 optimizer: adamw_bnb_8bit output_dir: miner_id_24 pad_to_sequence_len: true resume_from_checkpoint: null s2_attention: null sample_packing: false save_steps: 100 sequence_len: 2048 special_tokens: pad_token: <|eot_id|> strict: false tf32: true tokenizer_type: AutoTokenizer train_on_inputs: false trust_remote_code: true val_set_size: 0.04 wandb_entity: null wandb_mode: online wandb_name: 70c61e5c-5f68-4639-804d-89212f8c7742 wandb_project: Gradients-On-Demand wandb_run: your_name wandb_runid: 70c61e5c-5f68-4639-804d-89212f8c7742 warmup_steps: 10 weight_decay: 0.0 xformers_attention: null ``` </details> # f2b0fa58-8c68-4ee4-8ff5-65c996bc267d This model is a fine-tuned version of [llamafactory/tiny-random-Llama-3](https://huggingface.co/llamafactory/tiny-random-Llama-3) on the None dataset. It achieves the following results on the evaluation set: - Loss: 11.7233 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 32 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 10 - training_steps: 2123 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | 11.7614 | 0.0009 | 1 | 11.7621 | | 11.7349 | 0.0942 | 100 | 11.7364 | | 11.7389 | 0.1885 | 200 | 11.7346 | | 11.732 | 0.2827 | 300 | 11.7321 | | 11.7438 | 0.3769 | 400 | 11.7298 | | 11.7335 | 0.4711 | 500 | 11.7285 | | 11.7324 | 0.5654 | 600 | 11.7275 | | 11.7331 | 0.6596 | 700 | 11.7265 | | 11.7274 | 0.7538 | 800 | 11.7258 | | 11.7295 | 0.8481 | 900 | 11.7255 | | 11.731 | 0.9423 | 1000 | 11.7251 | | 13.0224 | 1.0365 | 1100 | 11.7249 | | 11.6643 | 1.1307 | 1200 | 11.7242 | | 11.7749 | 1.2250 | 1300 | 11.7240 | | 11.8175 | 1.3192 | 1400 | 11.7238 | | 10.3085 | 1.4134 | 1500 | 11.7236 | | 12.9048 | 1.5077 | 1600 | 11.7235 | | 10.6263 | 1.6019 | 1700 | 11.7234 | | 10.6834 | 1.6961 | 1800 | 11.7233 | | 11.5317 | 1.7903 | 1900 | 11.7233 | | 12.3242 | 1.8846 | 2000 | 11.7233 | | 10.9003 | 1.9788 | 2100 | 11.7233 | ### Framework versions - PEFT 0.13.2 - Transformers 4.46.0 - Pytorch 2.5.0+cu124 - Datasets 3.0.1 - Tokenizers 0.20.1

{"base_model": "llamafactory/tiny-random-Llama-3", "library_name": "peft", "license": "apache-2.0", "tags": ["axolotl", "generated_from_trainer"], "model-index": [{"name": "f2b0fa58-8c68-4ee4-8ff5-65c996bc267d", "results": []}]}

task

[ "TRANSLATION" ]

44,541

aimped/nlp-health-translation-base-en-fr

aimped

translation

[ "transformers", "pytorch", "marian", "text2text-generation", "medical", "translation", "medical translation", "en", "fr", "dataset:aimped/medical-translation-test-set", "arxiv:2407.12126", "license:cc-by-nc-4.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-10-26T08:55:01Z

2024-08-29T10:20:55+00:00

0

1

--- datasets: - aimped/medical-translation-test-set language: - en - fr license: cc-by-nc-4.0 metrics: - bleu pipeline_tag: translation tags: - medical - translation - medical translation extra_gated_heading: Access aimped/nlp-health-translation-base-en-fr on Hugging Face extra_gated_description: This is a form to enable access to this model on Hugging Face after you have been granted access from the Aimped. Please visit the [Aimped website](https://aimped.ai/) to Sign Up and accept our Terms of Use and Privacy Policy before submitting this form. Requests will be processed in 1-2 days. extra_gated_prompt: '**Your Hugging Face account email address MUST match the email you provide on the Aimped website or your request will not be approved.**' extra_gated_button_content: Submit extra_gated_fields: ? I agree to share my name, email address, and username with Aimped and confirm that I have already been granted download access on the Aimped website : checkbox widget: - text: 'Objective: Physical traumas are one of the important causes of mortality and morbidity in childhood. Permanent disabilities resulting from traumas constitute significant losses for the individual and society.' - text: Evidence is reported that a variety of chronic respiratory diseases, particularly COPD, asthma, bronchiectasis, lung cancer, interstitial lung diseases, and sarcoidosis, are significantly associated with poor clinical outcomes of COVID-19. --- <img src="https://raw.githubusercontent.com/ai-amplified/models/main/media/AimpedLogoDark.svg" alt="aimped logo" width="50%" height="50%"/> ### Description of the Model Paper: <a href="https://arxiv.org/abs/2407.12126" style="text-decoration: underline; color: blue;">LLMs-in-the-loop Part-1: Expert Small AI Models for Bio-Medical Text Translation</a> The Medical Translation AI model represents a specialized language model, trained for the accurate translations of medical documents from English to French. Its primary objective is to provide healthcare professionals, researchers, and individuals within the medical field with a reliable tool for the precise translation of a wide spectrum of medical documents.            The development of this model entailed the utilization of the <a href="https://github.com/Helsinki-NLP/OPUS-MT-train/tree/master/models/en-fr" style="text-decoration: underline; color: blue;">Hensinki/MarianMT</a> neural translation architecture, which required 2+ days of intensive training using A100 (24G RAM) GPU. To create an exceptionally high-quality corpus for training the translation model, we combined both publicly available and proprietary datasets. These datasets were further enriched by meticulously curated text collected from online sources. In addition, the inclusion of clinical and discharge reports from diverse healthcare institutions enhanced the dataset's depth and diversity. This meticulous curation process plays a pivotal role in ensuring the model's ability to generate accurate translations tailored specifically to the medical domain, meeting the stringent standards expected by our users. The versatility of the Medical Translation AI model extends to the translation of a wide array of healthcare-related documents, encompassing medical reports, patient records, medication instructions, research manuscripts, clinical trial documents, and more. By harnessing the capabilities of this model, users can efficiently and dependably obtain translations, thereby streamlining and expediting the often complex task of language translation within the medical field. The model we have developed outperforms leading translation companies like Google, Helsinki-Opus/MarianMT, and DeepL when compared against our meticulously curated proprietary test data set. <table style="border-collapse: collapse; width: 605px; height: 117px; border: 1px lightgray;"> <tbody> <tr> <td style="width: 19.5041%; border: 1px lightgray;"> </td> <td style="width: 20.6612%; text-align: center; border: 1px lightgray; font-size: 16px;">ROUGE</td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;">BLEU</td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;">METEOR</td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;">BERT</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">Aimped</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.85</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.62</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.83</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.95</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">Google</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.84</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.61</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.82</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.95</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">DeepL</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.81</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.57</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.78</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.94</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">Opus/MarianMT</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.80</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.52</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.76</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.93</td> </tr> </tbody> </table> ## Why should you use Aimped API? To get started, you can easily use our open-source version of the models for research purposes. However, the models provided through the Aimped API are trained on new data every three months. This ensures that the models understand ongoing healthcare developments in the world and can identify the most relevant medical terminology without a knowledge cutoff. In addition, we implement post/pre processing steps to improve the translation quality. Naturally, our quality control ensures that the models' performance always remains at least similar to previous versions. ## How to Use: To get the right results, use this function. - Install requirements ```python !pip install transformers !pip install sentencepiece !pip install aimped import nltk nltk.download('punkt') ``` - import libraries ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline from aimped.nlp.translation import text_translate import torch device = "cuda" if torch.cuda.is_available() else "cpu" ``` - load model ```python model_path = "aimped/nlp-health-translation-base-en-fr" tokenizer = AutoTokenizer.from_pretrained(model_path) model = AutoModelForSeq2SeqLM.from_pretrained(model_path) ``` ```python translater = pipeline( task="translation_en_to_fr", model=model, tokenizer=tokenizer, device= device, max_length=512, num_beams=7, early_stopping=False, num_return_sequences=1, do_sample=False, ) ``` - Use Model: ```python sentence = "Conclusion: According to our findings, the most common causes of major injuries in childhood are falls and home accidents." translated_text = text_translate([sentence],source_lang="en", pipeline=translater) ``` ## Test Set Trainin data: Public and in-house datasets. Test data: Public and in-house datasets which is available <a href="https://github.com/ai-amplified/models/tree/main/medical_translation/test_data/en-fr pairs">here</a>.

null

BioNLP

<img src="https://raw.githubusercontent.com/ai-amplified/models/main/media/AimpedLogoDark.svg" alt="aimped logo" width="50%" height="50%"/> ### Description of the Model Paper: <a href="https://arxiv.org/abs/2407.12126" style="text-decoration: underline; color: blue;">LLMs-in-the-loop Part-1: Expert Small AI Models for Bio-Medical Text Translation</a> The Medical Translation AI model represents a specialized language model, trained for the accurate translations of medical documents from English to French. Its primary objective is to provide healthcare professionals, researchers, and individuals within the medical field with a reliable tool for the precise translation of a wide spectrum of medical documents.            The development of this model entailed the utilization of the <a href="https://github.com/Helsinki-NLP/OPUS-MT-train/tree/master/models/en-fr" style="text-decoration: underline; color: blue;">Hensinki/MarianMT</a> neural translation architecture, which required 2+ days of intensive training using A100 (24G RAM) GPU. To create an exceptionally high-quality corpus for training the translation model, we combined both publicly available and proprietary datasets. These datasets were further enriched by meticulously curated text collected from online sources. In addition, the inclusion of clinical and discharge reports from diverse healthcare institutions enhanced the dataset's depth and diversity. This meticulous curation process plays a pivotal role in ensuring the model's ability to generate accurate translations tailored specifically to the medical domain, meeting the stringent standards expected by our users. The versatility of the Medical Translation AI model extends to the translation of a wide array of healthcare-related documents, encompassing medical reports, patient records, medication instructions, research manuscripts, clinical trial documents, and more. By harnessing the capabilities of this model, users can efficiently and dependably obtain translations, thereby streamlining and expediting the often complex task of language translation within the medical field. The model we have developed outperforms leading translation companies like Google, Helsinki-Opus/MarianMT, and DeepL when compared against our meticulously curated proprietary test data set. <table style="border-collapse: collapse; width: 605px; height: 117px; border: 1px lightgray;"> <tbody> <tr> <td style="width: 19.5041%; border: 1px lightgray;"> </td> <td style="width: 20.6612%; text-align: center; border: 1px lightgray; font-size: 16px;">ROUGE</td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;">BLEU</td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;">METEOR</td> <td style="width: 20%; text-align: center; border: 1px lightgray; font-size: 16px;">BERT</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">Aimped</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.85</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.62</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.83</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.95</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">Google</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.84</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.61</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.82</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.95</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">DeepL</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.81</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.57</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.78</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.94</td> </tr> <tr> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">Opus/MarianMT</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.80</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.52</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.76</td> <td style="text-align: center; border: 1px lightgray; font-size: 16px;">0.93</td> </tr> </tbody> </table> ## Why should you use Aimped API? To get started, you can easily use our open-source version of the models for research purposes. However, the models provided through the Aimped API are trained on new data every three months. This ensures that the models understand ongoing healthcare developments in the world and can identify the most relevant medical terminology without a knowledge cutoff. In addition, we implement post/pre processing steps to improve the translation quality. Naturally, our quality control ensures that the models' performance always remains at least similar to previous versions. ## How to Use: To get the right results, use this function. - Install requirements ```python !pip install transformers !pip install sentencepiece !pip install aimped import nltk nltk.download('punkt') ``` - import libraries ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline from aimped.nlp.translation import text_translate import torch device = "cuda" if torch.cuda.is_available() else "cpu" ``` - load model ```python model_path = "aimped/nlp-health-translation-base-en-fr" tokenizer = AutoTokenizer.from_pretrained(model_path) model = AutoModelForSeq2SeqLM.from_pretrained(model_path) ``` ```python translater = pipeline( task="translation_en_to_fr", model=model, tokenizer=tokenizer, device= device, max_length=512, num_beams=7, early_stopping=False, num_return_sequences=1, do_sample=False, ) ``` - Use Model: ```python sentence = "Conclusion: According to our findings, the most common causes of major injuries in childhood are falls and home accidents." translated_text = text_translate([sentence],source_lang="en", pipeline=translater) ``` ## Test Set Trainin data: Public and in-house datasets. Test data: Public and in-house datasets which is available <a href="https://github.com/ai-amplified/models/tree/main/medical_translation/test_data/en-fr pairs">here</a>.

{"datasets": ["aimped/medical-translation-test-set"], "language": ["en", "fr"], "license": "cc-by-nc-4.0", "metrics": ["bleu"], "pipeline_tag": "translation", "tags": ["medical", "translation", "medical translation"], "extra_gated_heading": "Access aimped/nlp-health-translation-base-en-fr on Hugging Face", "extra_gated_description": "This is a form to enable access to this model on Hugging Face after you have been granted access from the Aimped. Please visit the [Aimped website](https://aimped.ai/) to Sign Up and accept our Terms of Use and Privacy Policy before submitting this form. Requests will be processed in 1-2 days.", "extra_gated_prompt": "**Your Hugging Face account email address MUST match the email you provide on the Aimped website or your request will not be approved.**", "extra_gated_button_content": "Submit", "extra_gated_fields": {"I agree to share my name, email address, and username with Aimped and confirm that I have already been granted download access on the Aimped website": "checkbox"}, "widget": [{"text": "Objective: Physical traumas are one of the important causes of mortality and morbidity in childhood. Permanent disabilities resulting from traumas constitute significant losses for the individual and society."}, {"text": "Evidence is reported that a variety of chronic respiratory diseases, particularly COPD, asthma, bronchiectasis, lung cancer, interstitial lung diseases, and sarcoidosis, are significantly associated with poor clinical outcomes of COVID-19."}]}

task

[ "TRANSLATION" ]

44,542

PardisSzah/Persian_NER_parsbert

PardisSzah

token-classification

[ "transformers", "safetensors", "bert", "token-classification", "fa", "license:mit", "autotrain_compatible", "region:us" ]

2024-02-16T19:26:01Z

2024-10-21T12:57:51+00:00

22

0

--- language: fa license: mit pipeline_tag: token-classification inference: false --- # NER_ARMAN_parsbert This model is fine-tuned for Named Entity Recognition task. It has been fine-tuned on ARMAN Dataset, using the pretrained model [bert-base-parsbert-ner-uncased](https://huggingface.co/HooshvareLab/bert-base-parsbert-ner-uncased). ## Usage ```python def predict(input_text): nlp = pipeline("ner", model="PardisSzah/Persian_NER_parsbert") output_predictions = [] for sequence in input_text: result = nlp(sequence) output_predictions.append(result) return output_predictions text = [ "علی اکبری در روز شنبه به دیدن مادرش مریم حسنی رفت و بعد به بیمارستان ارتش سر زد" ] output = predict(text) print(output) # output: [[{'entity': 'B-person', 'score': 0.9998951, 'index': 1, 'word': 'علی', 'start': 0, 'end': 3}, {'entity': 'I-person', 'score': 0.9999027, 'index': 2, 'word': 'اکبری', 'start': 4, 'end': 9}, {'entity': 'B-person', 'score': 0.9998709, 'index': 9, 'word': 'مریم', 'start': 36, 'end': 40}, {'entity': 'I-person', 'score': 0.9996691, 'index': 10, 'word': 'حسنی', 'start': 41, 'end': 45}, {'entity': 'B-facility', 'score': 0.9561743, 'index': 15, 'word': 'بیمارستان', 'start': 59, 'end': 68}, {'entity': 'I-facility', 'score': 0.9976502, 'index': 16, 'word': 'ارتش', 'start': 69, 'end': 73}]]

null

Non_BioNLP

# NER_ARMAN_parsbert This model is fine-tuned for Named Entity Recognition task. It has been fine-tuned on ARMAN Dataset, using the pretrained model [bert-base-parsbert-ner-uncased](https://huggingface.co/HooshvareLab/bert-base-parsbert-ner-uncased). ## Usage ```python def predict(input_text): nlp = pipeline("ner", model="PardisSzah/Persian_NER_parsbert") output_predictions = [] for sequence in input_text: result = nlp(sequence) output_predictions.append(result) return output_predictions text = [ "علی اکبری در روز شنبه به دیدن مادرش مریم حسنی رفت و بعد به بیمارستان ارتش سر زد" ] output = predict(text) print(output) # output: [[{'entity': 'B-person', 'score': 0.9998951, 'index': 1, 'word': 'علی', 'start': 0, 'end': 3}, {'entity': 'I-person', 'score': 0.9999027, 'index': 2, 'word': 'اکبری', 'start': 4, 'end': 9}, {'entity': 'B-person', 'score': 0.9998709, 'index': 9, 'word': 'مریم', 'start': 36, 'end': 40}, {'entity': 'I-person', 'score': 0.9996691, 'index': 10, 'word': 'حسنی', 'start': 41, 'end': 45}, {'entity': 'B-facility', 'score': 0.9561743, 'index': 15, 'word': 'بیمارستان', 'start': 59, 'end': 68}, {'entity': 'I-facility', 'score': 0.9976502, 'index': 16, 'word': 'ارتش', 'start': 69, 'end': 73}]]

{"language": "fa", "license": "mit", "pipeline_tag": "token-classification", "inference": false}

task

[ "NAMED_ENTITY_RECOGNITION" ]

44,543

broyal/gemma-7b-AWQ

broyal

text-generation

[ "transformers", "safetensors", "gemma", "text-generation", "arxiv:2305.14314", "arxiv:2312.11805", "arxiv:2009.03300", "arxiv:1905.07830", "arxiv:1911.11641", "arxiv:1904.09728", "arxiv:1905.10044", "arxiv:1907.10641", "arxiv:1811.00937", "arxiv:1809.02789", "arxiv:1911.01547", "arxiv:1705.03551", "arxiv:2107.03374", "arxiv:2108.07732", "arxiv:2110.14168", "arxiv:2304.06364", "arxiv:2206.04615", "arxiv:1804.06876", "arxiv:2110.08193", "arxiv:2009.11462", "arxiv:2101.11718", "arxiv:1804.09301", "arxiv:2109.07958", "arxiv:2203.09509", "license:other", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "4-bit", "awq", "region:us" ]

2024-07-19T13:44:20Z

2024-07-19T15:18:23+00:00

77

0

--- library_name: transformers license: other license_name: gemma-terms-of-use license_link: https://ai.google.dev/gemma/terms tags: [] extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged-in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- AWQ quantized version of gemma-7b model. --- # Gemma Model Card **Model Page**: [Gemma](https://ai.google.dev/gemma/docs) This model card corresponds to the 7B base version of the Gemma model. You can also visit the model card of the [2B base model](https://huggingface.co/google/gemma-2b), [7B instruct model](https://huggingface.co/google/gemma-7b-it), and [2B instruct model](https://huggingface.co/google/gemma-2b-it). **Resources and Technical Documentation**: * [Responsible Generative AI Toolkit](https://ai.google.dev/responsible) * [Gemma on Kaggle](https://www.kaggle.com/models/google/gemma) * [Gemma on Vertex Model Garden](https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335?version=gemma-7b-gg-hf) **Terms of Use**: [Terms](https://www.kaggle.com/models/google/gemma/license/consent) **Authors**: Google ## Model Information Summary description and brief definition of inputs and outputs. ### Description Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. They are text-to-text, decoder-only large language models, available in English, with open weights, pre-trained variants, and instruction-tuned variants. Gemma models are well-suited for a variety of text generation tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as a laptop, desktop or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. ### Usage Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase. #### Fine-tuning examples You can find fine-tuning notebooks under the [`examples/` directory](https://huggingface.co/google/gemma-7b/tree/main/examples). We provide: * A script to perform Supervised Fine-Tuning (SFT) on UltraChat dataset using [QLoRA](https://huggingface.co/papers/2305.14314) * A script to perform SFT using FSDP on TPU devices * A notebook that you can run on a free-tier Google Colab instance to perform SFT on English quotes dataset. You can also find the copy of the notebook [here](https://github.com/huggingface/notebooks/blob/main/peft/gemma_7b_english_quotes.ipynb). #### Running the model on a CPU ```python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a GPU using different precisions * _Using `torch.float16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.float16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using `torch.bfloat16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.bfloat16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` * _Using 8-bit precision (int8)_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using 4-bit precision_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Other optimizations * _Flash Attention 2_ First make sure to install `flash-attn` in your environment `pip install flash-attn` ```diff model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, + attn_implementation="flash_attention_2" ).to(0) ``` ### Inputs and outputs * **Input:** Text string, such as a question, a prompt, or a document to be summarized. * **Output:** Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ## Model Data Data used for model training and how the data was processed. ### Training Dataset These models were trained on a dataset of text data that includes a wide variety of sources, totaling 6 trillion tokens. Here are the key components: * Web Documents: A diverse collection of web text ensures the model is exposed to a broad range of linguistic styles, topics, and vocabulary. Primarily English-language content. * Code: Exposing the model to code helps it to learn the syntax and patterns of programming languages, which improves its ability to generate code or understand code-related questions. * Mathematics: Training on mathematical text helps the model learn logical reasoning, symbolic representation, and to address mathematical queries. The combination of these diverse data sources is crucial for training a powerful language model that can handle a wide variety of different tasks and text formats. ### Data Preprocessing Here are the key data cleaning and filtering methods applied to the training data: * CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering was applied at multiple stages in the data preparation process to ensure the exclusion of harmful and illegal content * Sensitive Data Filtering: As part of making Gemma pre-trained models safe and reliable, automated techniques were used to filter out certain personal information and other sensitive data from training sets. * Additional methods: Filtering based on content quality and safely in line with [our policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11). ## Implementation Information Details about the model internals. ### Hardware Gemma was trained using the latest generation of [Tensor Processing Unit (TPU)](https://cloud.google.com/tpu/docs/intro-to-tpu) hardware (TPUv5e). Training large language models requires significant computational power. TPUs, designed specifically for matrix operations common in machine learning, offer several advantages in this domain: * Performance: TPUs are specifically designed to handle the massive computations involved in training LLMs. They can speed up training considerably compared to CPUs. * Memory: TPUs often come with large amounts of high-bandwidth memory, allowing for the handling of large models and batch sizes during training. This can lead to better model quality. * Scalability: TPU Pods (large clusters of TPUs) provide a scalable solution for handling the growing complexity of large foundation models. You can distribute training across multiple TPU devices for faster and more efficient processing. * Cost-effectiveness: In many scenarios, TPUs can provide a more cost-effective solution for training large models compared to CPU-based infrastructure, especially when considering the time and resources saved due to faster training. * These advantages are aligned with [Google's commitments to operate sustainably](https://sustainability.google/operating-sustainably/). ### Software Training was done using [JAX](https://github.com/google/jax) and [ML Pathways](https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture). JAX allows researchers to take advantage of the latest generation of hardware, including TPUs, for faster and more efficient training of large models. ML Pathways is Google's latest effort to build artificially intelligent systems capable of generalizing across multiple tasks. This is specially suitable for [foundation models](https://ai.google/discover/foundation-models/), including large language models like these ones. Together, JAX and ML Pathways are used as described in the [paper about the Gemini family of models](https://arxiv.org/abs/2312.11805); "the 'single controller' programming model of Jax and Pathways allows a single Python process to orchestrate the entire training run, dramatically simplifying the development workflow." ## Evaluation Model evaluation metrics and results. ### Benchmark Results These models were evaluated against a large collection of different datasets and metrics to cover different aspects of text generation: | Benchmark | Metric | 2B Params | 7B Params | | ------------------------------ | ------------- | ----------- | --------- | | [MMLU](https://arxiv.org/abs/2009.03300) | 5-shot, top-1 | 42.3 | 64.3 | | [HellaSwag](https://arxiv.org/abs/1905.07830) | 0-shot |71.4 | 81.2 | | [PIQA](https://arxiv.org/abs/1911.11641) | 0-shot | 77.3 | 81.2 | | [SocialIQA](https://arxiv.org/abs/1904.09728) | 0-shot | 59.7 | 51.8 | | [BooIQ](https://arxiv.org/abs/1905.10044) | 0-shot | 69.4 | 83.2 | | [WinoGrande](https://arxiv.org/abs/1907.10641) | partial score | 65.4 | 72.3 | | [CommonsenseQA](https://arxiv.org/abs/1811.00937) | 7-shot | 65.3 | 71.3 | | [OpenBookQA](https://arxiv.org/abs/1809.02789) | | 47.8 | 52.8 | | [ARC-e](https://arxiv.org/abs/1911.01547) | | 73.2 | 81.5 | | [ARC-c](https://arxiv.org/abs/1911.01547) | | 42.1 | 53.2 | | [TriviaQA](https://arxiv.org/abs/1705.03551) | 5-shot | 53.2 | 63.4 | | [Natural Questions](https://github.com/google-research-datasets/natural-questions) | 5-shot | - | 23 | | [HumanEval](https://arxiv.org/abs/2107.03374) | pass@1 | 22.0 | 32.3 | | [MBPP](https://arxiv.org/abs/2108.07732) | 3-shot | 29.2 | 44.4 | | [GSM8K](https://arxiv.org/abs/2110.14168) | maj@1 | 17.7 | 46.4 | | [MATH](https://arxiv.org/abs/2108.07732) | 4-shot | 11.8 | 24.3 | | [AGIEval](https://arxiv.org/abs/2304.06364) | | 24.2 | 41.7 | | [BIG-Bench](https://arxiv.org/abs/2206.04615) | | 35.2 | 55.1 | | ------------------------------ | ------------- | ----------- | --------- | | **Average** | | **54.0** | **56.4** | ## Ethics and Safety Ethics and safety evaluation approach and results. ### Evaluation Approach Our evaluation methods include structured evaluations and internal red-teaming testing of relevant content policies. Red-teaming was conducted by a number of different teams, each with different goals and human evaluation metrics. These models were evaluated against a number of different categories relevant to ethics and safety, including: * Text-to-Text Content Safety: Human evaluation on prompts covering safety policies including child sexual abuse and exploitation, harassment, violence and gore, and hate speech. * Text-to-Text Representational Harms: Benchmark against relevant academic datasets such as [WinoBias](https://arxiv.org/abs/1804.06876) and [BBQ Dataset](https://arxiv.org/abs/2110.08193v2). * Memorization: Automated evaluation of memorization of training data, including the risk of personally identifiable information exposure. * Large-scale harm: Tests for "dangerous capabilities," such as chemical, biological, radiological, and nuclear (CBRN) risks. ### Evaluation Results The results of ethics and safety evaluations are within acceptable thresholds for meeting [internal policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11) for categories such as child safety, content safety, representational harms, memorization, large-scale harms. On top of robust internal evaluations, the results of well known safety benchmarks like BBQ, BOLD, Winogender, Winobias, RealToxicity, and TruthfulQA are shown here. | Benchmark | Metric | 2B Params | 7B Params | | ------------------------------ | ------------- | ----------- | --------- | | [RealToxicity](https://arxiv.org/abs/2009.11462) | average | 6.86 | 7.90 | | [BOLD](https://arxiv.org/abs/2101.11718) | | 45.57 | 49.08 | | [CrowS-Pairs](https://aclanthology.org/2020.emnlp-main.154/) | top-1 | 45.82 | 51.33 | | [BBQ Ambig](https://arxiv.org/abs/2110.08193v2) | 1-shot, top-1 | 62.58 | 92.54 | | [BBQ Disambig](https://arxiv.org/abs/2110.08193v2) | top-1 | 54.62 | 71.99 | | [Winogender](https://arxiv.org/abs/1804.09301) | top-1 | 51.25 | 54.17 | | [TruthfulQA](https://arxiv.org/abs/2109.07958) | | 44.84 | 31.81 | | [Winobias 1_2](https://arxiv.org/abs/1804.06876) | | 56.12 | 59.09 | | [Winobias 2_2](https://arxiv.org/abs/1804.06876) | | 91.10 | 92.23 | | [Toxigen](https://arxiv.org/abs/2203.09509) | | 29.77 | 39.59 | | ------------------------------ | ------------- | ----------- | --------- | ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit](http://ai.google.dev/gemma/responsible). * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques. ### Benefits At the time of release, this family of models provides high-performance open large language model implementations designed from the ground up for Responsible AI development compared to similarly sized models. Using the benchmark evaluation metrics described in this document, these models have shown to provide superior performance to other, comparably-sized open model alternatives.

null

Non_BioNLP

AWQ quantized version of gemma-7b model. --- # Gemma Model Card **Model Page**: [Gemma](https://ai.google.dev/gemma/docs) This model card corresponds to the 7B base version of the Gemma model. You can also visit the model card of the [2B base model](https://huggingface.co/google/gemma-2b), [7B instruct model](https://huggingface.co/google/gemma-7b-it), and [2B instruct model](https://huggingface.co/google/gemma-2b-it). **Resources and Technical Documentation**: * [Responsible Generative AI Toolkit](https://ai.google.dev/responsible) * [Gemma on Kaggle](https://www.kaggle.com/models/google/gemma) * [Gemma on Vertex Model Garden](https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335?version=gemma-7b-gg-hf) **Terms of Use**: [Terms](https://www.kaggle.com/models/google/gemma/license/consent) **Authors**: Google ## Model Information Summary description and brief definition of inputs and outputs. ### Description Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. They are text-to-text, decoder-only large language models, available in English, with open weights, pre-trained variants, and instruction-tuned variants. Gemma models are well-suited for a variety of text generation tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as a laptop, desktop or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. ### Usage Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase. #### Fine-tuning examples You can find fine-tuning notebooks under the [`examples/` directory](https://huggingface.co/google/gemma-7b/tree/main/examples). We provide: * A script to perform Supervised Fine-Tuning (SFT) on UltraChat dataset using [QLoRA](https://huggingface.co/papers/2305.14314) * A script to perform SFT using FSDP on TPU devices * A notebook that you can run on a free-tier Google Colab instance to perform SFT on English quotes dataset. You can also find the copy of the notebook [here](https://github.com/huggingface/notebooks/blob/main/peft/gemma_7b_english_quotes.ipynb). #### Running the model on a CPU ```python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto") input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a GPU using different precisions * _Using `torch.float16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.float16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using `torch.bfloat16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.bfloat16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` * _Using 8-bit precision (int8)_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using 4-bit precision_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b") model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Other optimizations * _Flash Attention 2_ First make sure to install `flash-attn` in your environment `pip install flash-attn` ```diff model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, + attn_implementation="flash_attention_2" ).to(0) ``` ### Inputs and outputs * **Input:** Text string, such as a question, a prompt, or a document to be summarized. * **Output:** Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ## Model Data Data used for model training and how the data was processed. ### Training Dataset These models were trained on a dataset of text data that includes a wide variety of sources, totaling 6 trillion tokens. Here are the key components: * Web Documents: A diverse collection of web text ensures the model is exposed to a broad range of linguistic styles, topics, and vocabulary. Primarily English-language content. * Code: Exposing the model to code helps it to learn the syntax and patterns of programming languages, which improves its ability to generate code or understand code-related questions. * Mathematics: Training on mathematical text helps the model learn logical reasoning, symbolic representation, and to address mathematical queries. The combination of these diverse data sources is crucial for training a powerful language model that can handle a wide variety of different tasks and text formats. ### Data Preprocessing Here are the key data cleaning and filtering methods applied to the training data: * CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering was applied at multiple stages in the data preparation process to ensure the exclusion of harmful and illegal content * Sensitive Data Filtering: As part of making Gemma pre-trained models safe and reliable, automated techniques were used to filter out certain personal information and other sensitive data from training sets. * Additional methods: Filtering based on content quality and safely in line with [our policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11). ## Implementation Information Details about the model internals. ### Hardware Gemma was trained using the latest generation of [Tensor Processing Unit (TPU)](https://cloud.google.com/tpu/docs/intro-to-tpu) hardware (TPUv5e). Training large language models requires significant computational power. TPUs, designed specifically for matrix operations common in machine learning, offer several advantages in this domain: * Performance: TPUs are specifically designed to handle the massive computations involved in training LLMs. They can speed up training considerably compared to CPUs. * Memory: TPUs often come with large amounts of high-bandwidth memory, allowing for the handling of large models and batch sizes during training. This can lead to better model quality. * Scalability: TPU Pods (large clusters of TPUs) provide a scalable solution for handling the growing complexity of large foundation models. You can distribute training across multiple TPU devices for faster and more efficient processing. * Cost-effectiveness: In many scenarios, TPUs can provide a more cost-effective solution for training large models compared to CPU-based infrastructure, especially when considering the time and resources saved due to faster training. * These advantages are aligned with [Google's commitments to operate sustainably](https://sustainability.google/operating-sustainably/). ### Software Training was done using [JAX](https://github.com/google/jax) and [ML Pathways](https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture). JAX allows researchers to take advantage of the latest generation of hardware, including TPUs, for faster and more efficient training of large models. ML Pathways is Google's latest effort to build artificially intelligent systems capable of generalizing across multiple tasks. This is specially suitable for [foundation models](https://ai.google/discover/foundation-models/), including large language models like these ones. Together, JAX and ML Pathways are used as described in the [paper about the Gemini family of models](https://arxiv.org/abs/2312.11805); "the 'single controller' programming model of Jax and Pathways allows a single Python process to orchestrate the entire training run, dramatically simplifying the development workflow." ## Evaluation Model evaluation metrics and results. ### Benchmark Results These models were evaluated against a large collection of different datasets and metrics to cover different aspects of text generation: | Benchmark | Metric | 2B Params | 7B Params | | ------------------------------ | ------------- | ----------- | --------- | | [MMLU](https://arxiv.org/abs/2009.03300) | 5-shot, top-1 | 42.3 | 64.3 | | [HellaSwag](https://arxiv.org/abs/1905.07830) | 0-shot |71.4 | 81.2 | | [PIQA](https://arxiv.org/abs/1911.11641) | 0-shot | 77.3 | 81.2 | | [SocialIQA](https://arxiv.org/abs/1904.09728) | 0-shot | 59.7 | 51.8 | | [BooIQ](https://arxiv.org/abs/1905.10044) | 0-shot | 69.4 | 83.2 | | [WinoGrande](https://arxiv.org/abs/1907.10641) | partial score | 65.4 | 72.3 | | [CommonsenseQA](https://arxiv.org/abs/1811.00937) | 7-shot | 65.3 | 71.3 | | [OpenBookQA](https://arxiv.org/abs/1809.02789) | | 47.8 | 52.8 | | [ARC-e](https://arxiv.org/abs/1911.01547) | | 73.2 | 81.5 | | [ARC-c](https://arxiv.org/abs/1911.01547) | | 42.1 | 53.2 | | [TriviaQA](https://arxiv.org/abs/1705.03551) | 5-shot | 53.2 | 63.4 | | [Natural Questions](https://github.com/google-research-datasets/natural-questions) | 5-shot | - | 23 | | [HumanEval](https://arxiv.org/abs/2107.03374) | pass@1 | 22.0 | 32.3 | | [MBPP](https://arxiv.org/abs/2108.07732) | 3-shot | 29.2 | 44.4 | | [GSM8K](https://arxiv.org/abs/2110.14168) | maj@1 | 17.7 | 46.4 | | [MATH](https://arxiv.org/abs/2108.07732) | 4-shot | 11.8 | 24.3 | | [AGIEval](https://arxiv.org/abs/2304.06364) | | 24.2 | 41.7 | | [BIG-Bench](https://arxiv.org/abs/2206.04615) | | 35.2 | 55.1 | | ------------------------------ | ------------- | ----------- | --------- | | **Average** | | **54.0** | **56.4** | ## Ethics and Safety Ethics and safety evaluation approach and results. ### Evaluation Approach Our evaluation methods include structured evaluations and internal red-teaming testing of relevant content policies. Red-teaming was conducted by a number of different teams, each with different goals and human evaluation metrics. These models were evaluated against a number of different categories relevant to ethics and safety, including: * Text-to-Text Content Safety: Human evaluation on prompts covering safety policies including child sexual abuse and exploitation, harassment, violence and gore, and hate speech. * Text-to-Text Representational Harms: Benchmark against relevant academic datasets such as [WinoBias](https://arxiv.org/abs/1804.06876) and [BBQ Dataset](https://arxiv.org/abs/2110.08193v2). * Memorization: Automated evaluation of memorization of training data, including the risk of personally identifiable information exposure. * Large-scale harm: Tests for "dangerous capabilities," such as chemical, biological, radiological, and nuclear (CBRN) risks. ### Evaluation Results The results of ethics and safety evaluations are within acceptable thresholds for meeting [internal policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11) for categories such as child safety, content safety, representational harms, memorization, large-scale harms. On top of robust internal evaluations, the results of well known safety benchmarks like BBQ, BOLD, Winogender, Winobias, RealToxicity, and TruthfulQA are shown here. | Benchmark | Metric | 2B Params | 7B Params | | ------------------------------ | ------------- | ----------- | --------- | | [RealToxicity](https://arxiv.org/abs/2009.11462) | average | 6.86 | 7.90 | | [BOLD](https://arxiv.org/abs/2101.11718) | | 45.57 | 49.08 | | [CrowS-Pairs](https://aclanthology.org/2020.emnlp-main.154/) | top-1 | 45.82 | 51.33 | | [BBQ Ambig](https://arxiv.org/abs/2110.08193v2) | 1-shot, top-1 | 62.58 | 92.54 | | [BBQ Disambig](https://arxiv.org/abs/2110.08193v2) | top-1 | 54.62 | 71.99 | | [Winogender](https://arxiv.org/abs/1804.09301) | top-1 | 51.25 | 54.17 | | [TruthfulQA](https://arxiv.org/abs/2109.07958) | | 44.84 | 31.81 | | [Winobias 1_2](https://arxiv.org/abs/1804.06876) | | 56.12 | 59.09 | | [Winobias 2_2](https://arxiv.org/abs/1804.06876) | | 91.10 | 92.23 | | [Toxigen](https://arxiv.org/abs/2203.09509) | | 29.77 | 39.59 | | ------------------------------ | ------------- | ----------- | --------- | ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit](http://ai.google.dev/gemma/responsible). * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques. ### Benefits At the time of release, this family of models provides high-performance open large language model implementations designed from the ground up for Responsible AI development compared to similarly sized models. Using the benchmark evaluation metrics described in this document, these models have shown to provide superior performance to other, comparably-sized open model alternatives.

{"library_name": "transformers", "license": "other", "license_name": "gemma-terms-of-use", "license_link": "https://ai.google.dev/gemma/terms", "tags": [], "extra_gated_heading": "Access Gemma on Hugging Face", "extra_gated_prompt": "To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged-in to Hugging Face and click below. Requests are processed immediately.", "extra_gated_button_content": "Acknowledge license"}

task

[ "QUESTION_ANSWERING", "SUMMARIZATION" ]

44,544

junzai/demotest

junzai

text-classification

[ "transformers", "pytorch", "bert", "text-classification", "generated_from_trainer", "en", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:05Z

2022-02-23T07:51:36+00:00

120

0

--- datasets: - glue language: - en license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: bert_finetuning_test results: - task: type: text-classification name: Text Classification dataset: name: GLUE MRPC type: glue args: mrpc metrics: - type: accuracy value: 0.8284313725490197 name: Accuracy - type: f1 value: 0.8817567567567567 name: F1 ---  # bert_finetuning_test This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.4023 - Accuracy: 0.8284 - F1: 0.8818 - Combined Score: 0.8551 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1.0 ### Training results ### Framework versions - Transformers 4.16.0.dev0 - Pytorch 1.10.1+cu102 - Datasets 1.17.0 - Tokenizers 0.11.0

null

Non_BioNLP

# bert_finetuning_test This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.4023 - Accuracy: 0.8284 - F1: 0.8818 - Combined Score: 0.8551 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1.0 ### Training results ### Framework versions - Transformers 4.16.0.dev0 - Pytorch 1.10.1+cu102 - Datasets 1.17.0 - Tokenizers 0.11.0

{"datasets": ["glue"], "language": ["en"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert_finetuning_test", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "GLUE MRPC", "type": "glue", "args": "mrpc"}, "metrics": [{"type": "accuracy", "value": 0.8284313725490197, "name": "Accuracy"}, {"type": "f1", "value": 0.8817567567567567, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,545

skrh/bert_finetuning-sentiment-model-all-samples-normal

skrh

text-classification

[ "transformers", "tensorboard", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "dataset:imdb", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-11-23T14:26:40Z

2023-11-23T16:31:33+00:00

5

0

--- base_model: distilbert-base-uncased datasets: - imdb license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: bert_finetuning-sentiment-model-all-samples-normal results: - task: type: text-classification name: Text Classification dataset: name: imdb type: imdb config: plain_text split: test args: plain_text metrics: - type: accuracy value: 0.93212 name: Accuracy - type: f1 value: 0.9323985181054057 name: F1 ---  # bert_finetuning-sentiment-model-all-samples-normal This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2325 - Accuracy: 0.9321 - F1: 0.9324 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.35.2 - Pytorch 2.1.0+cu118 - Datasets 2.15.0 - Tokenizers 0.15.0

null

Non_BioNLP

# bert_finetuning-sentiment-model-all-samples-normal This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2325 - Accuracy: 0.9321 - F1: 0.9324 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.35.2 - Pytorch 2.1.0+cu118 - Datasets 2.15.0 - Tokenizers 0.15.0

{"base_model": "distilbert-base-uncased", "datasets": ["imdb"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert_finetuning-sentiment-model-all-samples-normal", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "imdb", "type": "imdb", "config": "plain_text", "split": "test", "args": "plain_text"}, "metrics": [{"type": "accuracy", "value": 0.93212, "name": "Accuracy"}, {"type": "f1", "value": 0.9323985181054057, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,546

YakovElm/Hyperledger15SetFitModel_balance_ratio_Half

YakovElm

text-classification

[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

2023-06-01T14:27:59Z

2023-06-01T14:28:34+00:00

10

0

--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- # YakovElm/Hyperledger15SetFitModel_balance_ratio_Half This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("YakovElm/Hyperledger15SetFitModel_balance_ratio_Half") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

# YakovElm/Hyperledger15SetFitModel_balance_ratio_Half This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("YakovElm/Hyperledger15SetFitModel_balance_ratio_Half") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}

task

[ "TEXT_CLASSIFICATION" ]

44,547

navjordj/snl-large-summarization

navjordj

text2text-generation

[ "transformers", "pytorch", "t5", "text2text-generation", "generated_from_trainer", "dataset:navjordj/SNL_summarization", "license:apache-2.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2023-02-07T13:41:51Z

2023-02-08T12:02:00+00:00

19

1

--- datasets: - navjordj/SNL_summarization license: apache-2.0 tags: - generated_from_trainer inference: parameters: max_length: 160 model-index: - name: snl-large-summarization results: [] ---  # snl-large-summarization This model is a fine-tuned version of [north/t5_large_NCC_lm](https://huggingface.co/north/t5_large_NCC_lm) on the navjordj/SNL_summarization dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 32 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10.0 ### Training results ### Framework versions - Transformers 4.27.0.dev0 - Pytorch 1.13.1 - Datasets 2.9.0 - Tokenizers 0.13.2

null

Non_BioNLP

# snl-large-summarization This model is a fine-tuned version of [north/t5_large_NCC_lm](https://huggingface.co/north/t5_large_NCC_lm) on the navjordj/SNL_summarization dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 32 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10.0 ### Training results ### Framework versions - Transformers 4.27.0.dev0 - Pytorch 1.13.1 - Datasets 2.9.0 - Tokenizers 0.13.2

{"datasets": ["navjordj/SNL_summarization"], "license": "apache-2.0", "tags": ["generated_from_trainer"], "inference": {"parameters": {"max_length": 160}}, "model-index": [{"name": "snl-large-summarization", "results": []}]}

task

[ "SUMMARIZATION" ]

44,548

Shivam098/opt-translation

Shivam098

text2text-generation

[ "transformers", "pytorch", "safetensors", "mbart", "text2text-generation", "generated_from_trainer", "dataset:opus100", "base_model:Shivam098/opt-translation", "base_model:quantized:Shivam098/opt-translation", "autotrain_compatible", "endpoints_compatible", "4-bit", "gptq", "region:us" ]

2023-08-29T07:52:03Z

2023-09-05T06:46:59+00:00

53

0

--- base_model: Shivam098/opt-translation datasets: - opus100 tags: - generated_from_trainer model-index: - name: opt-Translation results: [] ---  # opt-Translation This model is a fine-tuned version of [Shivam098/opt-translation](https://huggingface.co/Shivam098/opt-translation) on the opus100 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 300 ### Training results ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

null

Non_BioNLP

# opt-Translation This model is a fine-tuned version of [Shivam098/opt-translation](https://huggingface.co/Shivam098/opt-translation) on the opus100 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 300 ### Training results ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

{"base_model": "Shivam098/opt-translation", "datasets": ["opus100"], "tags": ["generated_from_trainer"], "model-index": [{"name": "opt-Translation", "results": []}]}

task

[ "TRANSLATION" ]

44,549

RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf

RichardErkhov

null

[ "gguf", "endpoints_compatible", "region:us" ]

2024-05-31T01:19:56Z

2024-05-31T05:35:50+00:00

4

0

--- {} --- Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) Gugugo-koen-7B-V1.1 - GGUF - Model creator: https://huggingface.co/squarelike/ - Original model: https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1/ | Name | Quant method | Size | | ---- | ---- | ---- | | [Gugugo-koen-7B-V1.1.Q2_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q2_K.gguf) | Q2_K | 2.42GB | | [Gugugo-koen-7B-V1.1.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_XS.gguf) | IQ3_XS | 2.67GB | | [Gugugo-koen-7B-V1.1.IQ3_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_S.gguf) | IQ3_S | 2.81GB | | [Gugugo-koen-7B-V1.1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_S.gguf) | Q3_K_S | 2.81GB | | [Gugugo-koen-7B-V1.1.IQ3_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_M.gguf) | IQ3_M | 2.97GB | | [Gugugo-koen-7B-V1.1.Q3_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K.gguf) | Q3_K | 3.14GB | | [Gugugo-koen-7B-V1.1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_M.gguf) | Q3_K_M | 3.14GB | | [Gugugo-koen-7B-V1.1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_L.gguf) | Q3_K_L | 3.42GB | | [Gugugo-koen-7B-V1.1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_XS.gguf) | IQ4_XS | 3.47GB | | [Gugugo-koen-7B-V1.1.Q4_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_0.gguf) | Q4_0 | 3.64GB | | [Gugugo-koen-7B-V1.1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_NL.gguf) | IQ4_NL | 3.66GB | | [Gugugo-koen-7B-V1.1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_S.gguf) | Q4_K_S | 3.67GB | | [Gugugo-koen-7B-V1.1.Q4_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K.gguf) | Q4_K | 3.88GB | | [Gugugo-koen-7B-V1.1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_M.gguf) | Q4_K_M | 3.88GB | | [Gugugo-koen-7B-V1.1.Q4_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_1.gguf) | Q4_1 | 4.03GB | | [Gugugo-koen-7B-V1.1.Q5_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_0.gguf) | Q5_0 | 4.42GB | | [Gugugo-koen-7B-V1.1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_S.gguf) | Q5_K_S | 4.42GB | | [Gugugo-koen-7B-V1.1.Q5_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K.gguf) | Q5_K | 2.16GB | | [Gugugo-koen-7B-V1.1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_M.gguf) | Q5_K_M | 0.65GB | | [Gugugo-koen-7B-V1.1.Q5_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_1.gguf) | Q5_1 | 0.62GB | | [Gugugo-koen-7B-V1.1.Q6_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q6_K.gguf) | Q6_K | 0.56GB | | [Gugugo-koen-7B-V1.1.Q8_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q8_0.gguf) | Q8_0 | 0.63GB | Original model description: --- license: apache-2.0 datasets: - squarelike/sharegpt_deepl_ko_translation language: - en - ko pipeline_tag: translation --- # Gugugo-koen-7B-V1.1 Detail repo: [https://github.com/jwj7140/Gugugo](https://github.com/jwj7140/Gugugo) ![Gugugo](./logo.png) **Base Model**: [Llama-2-ko-7b](https://huggingface.co/beomi/llama-2-ko-7b) **Training Dataset**: [sharegpt_deepl_ko_translation](https://huggingface.co/datasets/squarelike/sharegpt_deepl_ko_translation). I trained with 1x A6000 GPUs for 90 hours. ## **Prompt Template** **KO->EN** ``` ### 한국어: {sentence}</끝> ### 영어: ``` **EN->KO** ``` ### 영어: {sentence}</끝> ### 한국어: ``` There are GPTQ, AWQ, and GGUF support. [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF) ## **Implementation Code** ```python from transformers import AutoModelForCausalLM, AutoTokenizer, StoppingCriteria, StoppingCriteriaList import torch repo = "squarelike/Gugugo-koen-7B-V1.1" model = AutoModelForCausalLM.from_pretrained( repo, load_in_4bit=True device_map='auto' ) tokenizer = AutoTokenizer.from_pretrained(repo) class StoppingCriteriaSub(StoppingCriteria): def __init__(self, stops = [], encounters=1): super().__init__() self.stops = [stop for stop in stops] def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor): for stop in self.stops: if torch.all((stop == input_ids[0][-len(stop):])).item(): return True return False stop_words_ids = torch.tensor([[829, 45107, 29958], [1533, 45107, 29958], [829, 45107, 29958], [21106, 45107, 29958]]).to("cuda") stopping_criteria = StoppingCriteriaList([StoppingCriteriaSub(stops=stop_words_ids)]) def gen(lan="en", x=""): if (lan == "ko"): prompt = f"### 한국어: {x}</끝>\n### 영어:" else: prompt = f"### 영어: {x}</끝>\n### 한국어:" gened = model.generate( **tokenizer( prompt, return_tensors='pt', return_token_type_ids=False ).to("cuda"), max_new_tokens=2000, temperature=0.3, # no_repeat_ngram_size=5, num_beams=5, stopping_criteria=stopping_criteria ) return tokenizer.decode(gened[0][1:]).replace(prompt+" ", "").replace("</끝>", "") print(gen(lan="en", x="Hello, world!")) ```

null

Non_BioNLP

Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) Gugugo-koen-7B-V1.1 - GGUF - Model creator: https://huggingface.co/squarelike/ - Original model: https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1/ | Name | Quant method | Size | | ---- | ---- | ---- | | [Gugugo-koen-7B-V1.1.Q2_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q2_K.gguf) | Q2_K | 2.42GB | | [Gugugo-koen-7B-V1.1.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_XS.gguf) | IQ3_XS | 2.67GB | | [Gugugo-koen-7B-V1.1.IQ3_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_S.gguf) | IQ3_S | 2.81GB | | [Gugugo-koen-7B-V1.1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_S.gguf) | Q3_K_S | 2.81GB | | [Gugugo-koen-7B-V1.1.IQ3_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ3_M.gguf) | IQ3_M | 2.97GB | | [Gugugo-koen-7B-V1.1.Q3_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K.gguf) | Q3_K | 3.14GB | | [Gugugo-koen-7B-V1.1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_M.gguf) | Q3_K_M | 3.14GB | | [Gugugo-koen-7B-V1.1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q3_K_L.gguf) | Q3_K_L | 3.42GB | | [Gugugo-koen-7B-V1.1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_XS.gguf) | IQ4_XS | 3.47GB | | [Gugugo-koen-7B-V1.1.Q4_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_0.gguf) | Q4_0 | 3.64GB | | [Gugugo-koen-7B-V1.1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.IQ4_NL.gguf) | IQ4_NL | 3.66GB | | [Gugugo-koen-7B-V1.1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_S.gguf) | Q4_K_S | 3.67GB | | [Gugugo-koen-7B-V1.1.Q4_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K.gguf) | Q4_K | 3.88GB | | [Gugugo-koen-7B-V1.1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_K_M.gguf) | Q4_K_M | 3.88GB | | [Gugugo-koen-7B-V1.1.Q4_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q4_1.gguf) | Q4_1 | 4.03GB | | [Gugugo-koen-7B-V1.1.Q5_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_0.gguf) | Q5_0 | 4.42GB | | [Gugugo-koen-7B-V1.1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_S.gguf) | Q5_K_S | 4.42GB | | [Gugugo-koen-7B-V1.1.Q5_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K.gguf) | Q5_K | 2.16GB | | [Gugugo-koen-7B-V1.1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_K_M.gguf) | Q5_K_M | 0.65GB | | [Gugugo-koen-7B-V1.1.Q5_1.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q5_1.gguf) | Q5_1 | 0.62GB | | [Gugugo-koen-7B-V1.1.Q6_K.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q6_K.gguf) | Q6_K | 0.56GB | | [Gugugo-koen-7B-V1.1.Q8_0.gguf](https://huggingface.co/RichardErkhov/squarelike_-_Gugugo-koen-7B-V1.1-gguf/blob/main/Gugugo-koen-7B-V1.1.Q8_0.gguf) | Q8_0 | 0.63GB | Original model description: --- license: apache-2.0 datasets: - squarelike/sharegpt_deepl_ko_translation language: - en - ko pipeline_tag: translation --- # Gugugo-koen-7B-V1.1 Detail repo: [https://github.com/jwj7140/Gugugo](https://github.com/jwj7140/Gugugo) ![Gugugo](./logo.png) **Base Model**: [Llama-2-ko-7b](https://huggingface.co/beomi/llama-2-ko-7b) **Training Dataset**: [sharegpt_deepl_ko_translation](https://huggingface.co/datasets/squarelike/sharegpt_deepl_ko_translation). I trained with 1x A6000 GPUs for 90 hours. ## **Prompt Template** **KO->EN** ``` ### 한국어: {sentence}</끝> ### 영어: ``` **EN->KO** ``` ### 영어: {sentence}</끝> ### 한국어: ``` There are GPTQ, AWQ, and GGUF support. [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GPTQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-AWQ) [https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF](https://huggingface.co/squarelike/Gugugo-koen-7B-V1.1-GGUF) ## **Implementation Code** ```python from transformers import AutoModelForCausalLM, AutoTokenizer, StoppingCriteria, StoppingCriteriaList import torch repo = "squarelike/Gugugo-koen-7B-V1.1" model = AutoModelForCausalLM.from_pretrained( repo, load_in_4bit=True device_map='auto' ) tokenizer = AutoTokenizer.from_pretrained(repo) class StoppingCriteriaSub(StoppingCriteria): def __init__(self, stops = [], encounters=1): super().__init__() self.stops = [stop for stop in stops] def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor): for stop in self.stops: if torch.all((stop == input_ids[0][-len(stop):])).item(): return True return False stop_words_ids = torch.tensor([[829, 45107, 29958], [1533, 45107, 29958], [829, 45107, 29958], [21106, 45107, 29958]]).to("cuda") stopping_criteria = StoppingCriteriaList([StoppingCriteriaSub(stops=stop_words_ids)]) def gen(lan="en", x=""): if (lan == "ko"): prompt = f"### 한국어: {x}</끝>\n### 영어:" else: prompt = f"### 영어: {x}</끝>\n### 한국어:" gened = model.generate( **tokenizer( prompt, return_tensors='pt', return_token_type_ids=False ).to("cuda"), max_new_tokens=2000, temperature=0.3, # no_repeat_ngram_size=5, num_beams=5, stopping_criteria=stopping_criteria ) return tokenizer.decode(gened[0][1:]).replace(prompt+" ", "").replace("</끝>", "") print(gen(lan="en", x="Hello, world!")) ```

{}

task

[ "TRANSLATION" ]

44,550

pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1-sts

pritamdeka

sentence-similarity

[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:5749", "loss:CosineSimilarityLoss", "arxiv:1908.10084", "base_model:pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "base_model:finetune:pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

2024-07-17T19:02:37Z

2024-07-17T19:03:24+00:00

10

0

--- base_model: pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 datasets: [] language: [] library_name: sentence-transformers metrics: - pearson_cosine - spearman_cosine - pearson_manhattan - spearman_manhattan - pearson_euclidean - spearman_euclidean - pearson_dot - spearman_dot - pearson_max - spearman_max pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:5749 - loss:CosineSimilarityLoss widget: - source_sentence: আমি "... comoving মহাজাগতিক বিশ্ৰাম ফ্ৰেমৰ তুলনাত ... সিংহ নক্ষত্ৰমণ্ডলৰ ফালে কিছু 371 কিলোমিটাৰ প্ৰতি ছেকেণ্ডত" আগবাঢ়িছো. sentences: - বাস্কেটবল খেলুৱৈগৰাকীয়ে নিজৰ দলৰ হৈ পইণ্ট লাভ কৰিবলৈ ওলাইছে। - আন কোনো বস্তুৰ লগত আপেক্ষিক নহোৱা কোনো ‘ষ্টিল’ নাই। - এজনী ছোৱালীয়ে বতাহ বাদ্যযন্ত্ৰ বজায়। - source_sentence: চাৰিটা ল’ৰা-ছোৱালীয়ে ভঁৰালৰ জীৱ-জন্তুবোৰলৈ চাই আছে। sentences: - ডাইনিং টেবুল এখনৰ চাৰিওফালে বৃদ্ধৰ দল এটাই পোজ দিছে। - বিকিনি পিন্ধা চাৰিগৰাকী মহিলাই বিলত ভলীবল খেলি আছে। - ল’ৰা-ছোৱালীয়ে ভেড়া চাই। - source_sentence: ডালত বহি থকা দুটা টান ঈগল। sentences: - জাতৰ জেব্ৰা ডানিঅ’ অত্যন্ত কঠোৰ মাছ, ইহঁতক হত্যা কৰাটো প্ৰায় কঠিন। - এটা ডালত দুটা ঈগল বহি আছে। - নূন্যতম মজুৰিৰ আইনসমূহে কম দক্ষ, কম উৎপাদনশীল লোকক আটাইতকৈ বেছি আঘাত দিয়ে। - source_sentence: '"মই আচলতে যি বিচাৰিছো সেয়া হৈছে মুছলমান জনসংখ্যাৰ এটা অনুমান..." @ThanosK আৰু @T.E.D., এটা সামগ্ৰিক, সাধাৰণ জনসংখ্যাৰ অনুমান f.e.' sentences: - এগৰাকী মহিলাই সেউজীয়া পিঁয়াজ কাটি আছে। - তলত দিয়া কথাখিনি মোৰ কুকুৰ কাণৰ দৰে কপিৰ পৰা লোৱা হৈছে নিউ পেংগুইন এটলাছ অৱ মেডিভেল হিষ্ট্ৰীৰ। - আমাৰ দৰে সৌৰজগতৰ কোনো তাৰকাৰাজ্যৰ বাহিৰত থকাটো সম্ভৱ হ’ব পাৰে। - source_sentence: ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে। sentences: - গছৰ শাৰী এটাৰ সন্মুখত পথাৰত ভেড়া চৰিছে। - এজন মানুহে গীটাৰ বজাই আছে। - ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে। model-index: - name: SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 results: - task: type: semantic-similarity name: Semantic Similarity dataset: name: pritamdeka/stsb assamese translated dev type: pritamdeka/stsb-assamese-translated-dev metrics: - type: pearson_cosine value: 0.8525258323169252 name: Pearson Cosine - type: spearman_cosine value: 0.8506593647943235 name: Spearman Cosine - type: pearson_manhattan value: 0.8334889460288037 name: Pearson Manhattan - type: spearman_manhattan value: 0.843042040822402 name: Spearman Manhattan - type: pearson_euclidean value: 0.8351723933495433 name: Pearson Euclidean - type: spearman_euclidean value: 0.8450734552112781 name: Spearman Euclidean - type: pearson_dot value: 0.8273071926204811 name: Pearson Dot - type: spearman_dot value: 0.8277520425148079 name: Spearman Dot - type: pearson_max value: 0.8525258323169252 name: Pearson Max - type: spearman_max value: 0.8506593647943235 name: Spearman Max - task: type: semantic-similarity name: Semantic Similarity dataset: name: pritamdeka/stsb assamese translated test type: pritamdeka/stsb-assamese-translated-test metrics: - type: pearson_cosine value: 0.8138083526567048 name: Pearson Cosine - type: spearman_cosine value: 0.8119367763029309 name: Spearman Cosine - type: pearson_manhattan value: 0.8044112753419641 name: Pearson Manhattan - type: spearman_manhattan value: 0.8073243490029997 name: Spearman Manhattan - type: pearson_euclidean value: 0.805728285628756 name: Pearson Euclidean - type: spearman_euclidean value: 0.8086070843216111 name: Spearman Euclidean - type: pearson_dot value: 0.7754575809083841 name: Pearson Dot - type: spearman_dot value: 0.7720173359758135 name: Spearman Dot - type: pearson_max value: 0.8138083526567048 name: Pearson Max - type: spearman_max value: 0.8119367763029309 name: Spearman Max --- # SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 tokens - **Similarity Function:** Cosine Similarity    ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1-sts") # Run inference sentences = [ 'ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে।', 'ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে।', 'এজন মানুহে গীটাৰ বজাই আছে।', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-dev` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) | Metric | Value | |:--------------------|:-----------| | pearson_cosine | 0.8525 | | **spearman_cosine** | **0.8507** | | pearson_manhattan | 0.8335 | | spearman_manhattan | 0.843 | | pearson_euclidean | 0.8352 | | spearman_euclidean | 0.8451 | | pearson_dot | 0.8273 | | spearman_dot | 0.8278 | | pearson_max | 0.8525 | | spearman_max | 0.8507 | #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-test` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) | Metric | Value | |:--------------------|:-----------| | pearson_cosine | 0.8138 | | **spearman_cosine** | **0.8119** | | pearson_manhattan | 0.8044 | | spearman_manhattan | 0.8073 | | pearson_euclidean | 0.8057 | | spearman_euclidean | 0.8086 | | pearson_dot | 0.7755 | | spearman_dot | 0.772 | | pearson_max | 0.8138 | | spearman_max | 0.8119 |   ## Training Details ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `num_train_epochs`: 10 - `warmup_ratio`: 0.1 - `fp16`: True - `load_best_model_at_end`: True #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: True - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs | Epoch | Step | Training Loss | loss | pritamdeka/stsb-assamese-translated-dev_spearman_cosine | pritamdeka/stsb-assamese-translated-test_spearman_cosine | |:----------:|:-------:|:-------------:|:----------:|:-------------------------------------------------------:|:--------------------------------------------------------:| | 1.1111 | 100 | 0.0331 | 0.0259 | 0.8482 | - | | **2.2222** | **200** | **0.0176** | **0.0253** | **0.8515** | **-** | | 3.3333 | 300 | 0.011 | 0.0253 | 0.8513 | - | | 4.4444 | 400 | 0.0066 | 0.0259 | 0.8492 | - | | 5.5556 | 500 | 0.0048 | 0.0255 | 0.8511 | - | | 6.6667 | 600 | 0.0037 | 0.0256 | 0.8508 | - | | 7.7778 | 700 | 0.0033 | 0.0254 | 0.8515 | - | | 8.8889 | 800 | 0.0029 | 0.0255 | 0.8512 | - | | 10.0 | 900 | 0.0027 | 0.0257 | 0.8507 | 0.8119 | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.12 - Sentence Transformers: 3.0.1 - Transformers: 4.42.4 - PyTorch: 2.3.1+cu121 - Accelerate: 0.32.1 - Datasets: 2.20.0 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ```

null

Non_BioNLP

# SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1](https://huggingface.co/pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 tokens - **Similarity Function:** Cosine Similarity    ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1-sts") # Run inference sentences = [ 'ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে।', 'ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে।', 'এজন মানুহে গীটাৰ বজাই আছে।', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-dev` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) | Metric | Value | |:--------------------|:-----------| | pearson_cosine | 0.8525 | | **spearman_cosine** | **0.8507** | | pearson_manhattan | 0.8335 | | spearman_manhattan | 0.843 | | pearson_euclidean | 0.8352 | | spearman_euclidean | 0.8451 | | pearson_dot | 0.8273 | | spearman_dot | 0.8278 | | pearson_max | 0.8525 | | spearman_max | 0.8507 | #### Semantic Similarity * Dataset: `pritamdeka/stsb-assamese-translated-test` * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) | Metric | Value | |:--------------------|:-----------| | pearson_cosine | 0.8138 | | **spearman_cosine** | **0.8119** | | pearson_manhattan | 0.8044 | | spearman_manhattan | 0.8073 | | pearson_euclidean | 0.8057 | | spearman_euclidean | 0.8086 | | pearson_dot | 0.7755 | | spearman_dot | 0.772 | | pearson_max | 0.8138 | | spearman_max | 0.8119 |   ## Training Details ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `num_train_epochs`: 10 - `warmup_ratio`: 0.1 - `fp16`: True - `load_best_model_at_end`: True #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 64 - `per_device_eval_batch_size`: 64 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `fp16`: True - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: None - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs | Epoch | Step | Training Loss | loss | pritamdeka/stsb-assamese-translated-dev_spearman_cosine | pritamdeka/stsb-assamese-translated-test_spearman_cosine | |:----------:|:-------:|:-------------:|:----------:|:-------------------------------------------------------:|:--------------------------------------------------------:| | 1.1111 | 100 | 0.0331 | 0.0259 | 0.8482 | - | | **2.2222** | **200** | **0.0176** | **0.0253** | **0.8515** | **-** | | 3.3333 | 300 | 0.011 | 0.0253 | 0.8513 | - | | 4.4444 | 400 | 0.0066 | 0.0259 | 0.8492 | - | | 5.5556 | 500 | 0.0048 | 0.0255 | 0.8511 | - | | 6.6667 | 600 | 0.0037 | 0.0256 | 0.8508 | - | | 7.7778 | 700 | 0.0033 | 0.0254 | 0.8515 | - | | 8.8889 | 800 | 0.0029 | 0.0255 | 0.8512 | - | | 10.0 | 900 | 0.0027 | 0.0257 | 0.8507 | 0.8119 | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.12 - Sentence Transformers: 3.0.1 - Transformers: 4.42.4 - PyTorch: 2.3.1+cu121 - Accelerate: 0.32.1 - Datasets: 2.20.0 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ```

{"base_model": "pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "datasets": [], "language": [], "library_name": "sentence-transformers", "metrics": ["pearson_cosine", "spearman_cosine", "pearson_manhattan", "spearman_manhattan", "pearson_euclidean", "spearman_euclidean", "pearson_dot", "spearman_dot", "pearson_max", "spearman_max"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:5749", "loss:CosineSimilarityLoss"], "widget": [{"source_sentence": "আমি \"... comoving মহাজাগতিক বিশ্ৰাম ফ্ৰেমৰ তুলনাত ... সিংহ নক্ষত্ৰমণ্ডলৰ ফালে কিছু 371 কিলোমিটাৰ প্ৰতি ছেকেণ্ডত\" আগবাঢ়িছো.", "sentences": ["বাস্কেটবল খেলুৱৈগৰাকীয়ে নিজৰ দলৰ হৈ পইণ্ট লাভ কৰিবলৈ ওলাইছে।", "আন কোনো বস্তুৰ লগত আপেক্ষিক নহোৱা কোনো ‘ষ্টিল’ নাই।", "এজনী ছোৱালীয়ে বতাহ বাদ্যযন্ত্ৰ বজায়।"]}, {"source_sentence": "চাৰিটা ল’ৰা-ছোৱালীয়ে ভঁৰালৰ জীৱ-জন্তুবোৰলৈ চাই আছে।", "sentences": ["ডাইনিং টেবুল এখনৰ চাৰিওফালে বৃদ্ধৰ দল এটাই পোজ দিছে।", "বিকিনি পিন্ধা চাৰিগৰাকী মহিলাই বিলত ভলীবল খেলি আছে।", "ল’ৰা-ছোৱালীয়ে ভেড়া চাই।"]}, {"source_sentence": "ডালত বহি থকা দুটা টান ঈগল।", "sentences": ["জাতৰ জেব্ৰা ডানিঅ’ অত্যন্ত কঠোৰ মাছ, ইহঁতক হত্যা কৰাটো প্ৰায় কঠিন।", "এটা ডালত দুটা ঈগল বহি আছে।", "নূন্যতম মজুৰিৰ আইনসমূহে কম দক্ষ, কম উৎপাদনশীল লোকক আটাইতকৈ বেছি আঘাত দিয়ে।"]}, {"source_sentence": "\"মই আচলতে যি বিচাৰিছো সেয়া হৈছে মুছলমান জনসংখ্যাৰ এটা অনুমান...\" @ThanosK আৰু @T.E.D., এটা সামগ্ৰিক, সাধাৰণ জনসংখ্যাৰ অনুমান f.e.", "sentences": ["এগৰাকী মহিলাই সেউজীয়া পিঁয়াজ কাটি আছে।", "তলত দিয়া কথাখিনি মোৰ কুকুৰ কাণৰ দৰে কপিৰ পৰা লোৱা হৈছে নিউ পেংগুইন এটলাছ অৱ মেডিভেল হিষ্ট্ৰীৰ।", "আমাৰ দৰে সৌৰজগতৰ কোনো তাৰকাৰাজ্যৰ বাহিৰত থকাটো সম্ভৱ হ’ব পাৰে।"]}, {"source_sentence": "ইণ্টাৰনেট কেমেৰাৰ জৰিয়তে এগৰাকী ছোৱালীৰ লগত কথা পাতিলে মানুহজনে।", "sentences": ["গছৰ শাৰী এটাৰ সন্মুখত পথাৰত ভেড়া চৰিছে।", "এজন মানুহে গীটাৰ বজাই আছে।", "ৱেবকেমৰ জৰিয়তে এগৰাকী ছোৱালীৰ সৈতে কথা পাতিছে এজন কিশোৰে।"]}], "model-index": [{"name": "SentenceTransformer based on pritamdeka/muril-base-cased-assamese-indicxnli-random-negatives-v1", "results": [{"task": {"type": "semantic-similarity", "name": "Semantic Similarity"}, "dataset": {"name": "pritamdeka/stsb assamese translated dev", "type": "pritamdeka/stsb-assamese-translated-dev"}, "metrics": [{"type": "pearson_cosine", "value": 0.8525258323169252, "name": "Pearson Cosine"}, {"type": "spearman_cosine", "value": 0.8506593647943235, "name": "Spearman Cosine"}, {"type": "pearson_manhattan", "value": 0.8334889460288037, "name": "Pearson Manhattan"}, {"type": "spearman_manhattan", "value": 0.843042040822402, "name": "Spearman Manhattan"}, {"type": "pearson_euclidean", "value": 0.8351723933495433, "name": "Pearson Euclidean"}, {"type": "spearman_euclidean", "value": 0.8450734552112781, "name": "Spearman Euclidean"}, {"type": "pearson_dot", "value": 0.8273071926204811, "name": "Pearson Dot"}, {"type": "spearman_dot", "value": 0.8277520425148079, "name": "Spearman Dot"}, {"type": "pearson_max", "value": 0.8525258323169252, "name": "Pearson Max"}, {"type": "spearman_max", "value": 0.8506593647943235, "name": "Spearman Max"}]}, {"task": {"type": "semantic-similarity", "name": "Semantic Similarity"}, "dataset": {"name": "pritamdeka/stsb assamese translated test", "type": "pritamdeka/stsb-assamese-translated-test"}, "metrics": [{"type": "pearson_cosine", "value": 0.8138083526567048, "name": "Pearson Cosine"}, {"type": "spearman_cosine", "value": 0.8119367763029309, "name": "Spearman Cosine"}, {"type": "pearson_manhattan", "value": 0.8044112753419641, "name": "Pearson Manhattan"}, {"type": "spearman_manhattan", "value": 0.8073243490029997, "name": "Spearman Manhattan"}, {"type": "pearson_euclidean", "value": 0.805728285628756, "name": "Pearson Euclidean"}, {"type": "spearman_euclidean", "value": 0.8086070843216111, "name": "Spearman Euclidean"}, {"type": "pearson_dot", "value": 0.7754575809083841, "name": "Pearson Dot"}, {"type": "spearman_dot", "value": 0.7720173359758135, "name": "Spearman Dot"}, {"type": "pearson_max", "value": 0.8138083526567048, "name": "Pearson Max"}, {"type": "spearman_max", "value": 0.8119367763029309, "name": "Spearman Max"}]}]}]}

task

[ "TEXT_CLASSIFICATION", "SEMANTIC_SIMILARITY" ]

44,551

luckyf1998/distilbert-base-uncased-finetuned-emotion

luckyf1998

text-classification

[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:emotion", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-08-03T11:31:49Z

2023-08-03T14:34:09+00:00

16

0

--- datasets: - emotion license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion args: split metrics: - type: accuracy value: 0.923 name: Accuracy - type: f1 value: 0.9227217081326218 name: F1 ---  # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2148 - Accuracy: 0.923 - F1: 0.9227 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 0.8269 | 1.0 | 250 | 0.3107 | 0.909 | 0.9072 | | 0.2402 | 2.0 | 500 | 0.2148 | 0.923 | 0.9227 | ### Framework versions - Transformers 4.16.2 - Pytorch 2.0.0 - Datasets 1.16.1 - Tokenizers 0.13.3

null

Non_BioNLP

# distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2148 - Accuracy: 0.923 - F1: 0.9227 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 0.8269 | 1.0 | 250 | 0.3107 | 0.909 | 0.9072 | | 0.2402 | 2.0 | 500 | 0.2148 | 0.923 | 0.9227 | ### Framework versions - Transformers 4.16.2 - Pytorch 2.0.0 - Datasets 1.16.1 - Tokenizers 0.13.3

{"datasets": ["emotion"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-emotion", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "args": "split"}, "metrics": [{"type": "accuracy", "value": 0.923, "name": "Accuracy"}, {"type": "f1", "value": 0.9227217081326218, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,552

Z3R6X/gpt4all_dpo_instruct

Z3R6X

text-generation

[ "transformers", "pytorch", "gptj", "text-generation", "en", "dataset:Dahoas/instruct-synthetic-prompt-responses", "arxiv:2305.18290", "license:cc-by-nc-4.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2024-03-10T15:27:19Z

2024-04-19T08:02:51+00:00

16

0

--- datasets: - Dahoas/instruct-synthetic-prompt-responses language: - en license: cc-by-nc-4.0 pipeline_tag: text-generation --- Question answering model finetuned from [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) with [Direct Preference Optimization](https://arxiv.org/abs/2305.18290). \ Dataset: [Dahoas/instruct-synthetic-prompt-responses](https://huggingface.co/datasets/Dahoas/instruct-synthetic-prompt-responses). The model was finetuned with the following promt: \ ``"Answer the following question in context:\n\nQuestion: " + samples["prompt"] + " Answer: "`` \ It should be benefical to use the same or a similar prompt for inference. An increase in performance compared to [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) was observed when using two-shot Chain-of-Thought prompting. | HellaSwag | WinoGrande | BooLQ | ARC-c | |:------:|:------:|:------:|:------:| | 62.37% | 63.3% | 65.2% | 32.76% |

null

Non_BioNLP

Question answering model finetuned from [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) with [Direct Preference Optimization](https://arxiv.org/abs/2305.18290). \ Dataset: [Dahoas/instruct-synthetic-prompt-responses](https://huggingface.co/datasets/Dahoas/instruct-synthetic-prompt-responses). The model was finetuned with the following promt: \ ``"Answer the following question in context:\n\nQuestion: " + samples["prompt"] + " Answer: "`` \ It should be benefical to use the same or a similar prompt for inference. An increase in performance compared to [GPT4All-J v1.3](https://huggingface.co/nomic-ai/gpt4all-j) was observed when using two-shot Chain-of-Thought prompting. | HellaSwag | WinoGrande | BooLQ | ARC-c | |:------:|:------:|:------:|:------:| | 62.37% | 63.3% | 65.2% | 32.76% |

{"datasets": ["Dahoas/instruct-synthetic-prompt-responses"], "language": ["en"], "license": "cc-by-nc-4.0", "pipeline_tag": "text-generation"}

task

[ "QUESTION_ANSWERING" ]

44,553

emozilla/LLongMA-2-13b-storysummarizer-GGML

emozilla

null

[ "dataset:emozilla/booksum-summary-analysis_llama-8192", "dataset:kmfoda/booksum", "license:apache-2.0", "region:us" ]

2023-08-07T15:28:28Z

2023-08-07T18:51:05+00:00

0

--- datasets: - emozilla/booksum-summary-analysis_llama-8192 - kmfoda/booksum license: apache-2.0 --- GGML version of [emozilla/LLongMA-2-13b-storysummarizer](https://huggingface.co/emozilla/LLongMA-2-13b-storysummarizer). Use `n_ctx=8192` and `rope_freq_scale=0.5`. For an excellent user interface, check out [LM Studio](https://lmstudio.ai/). # LLongMA-2-13b-storysummarizer [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl) This is a fine-tuned version of [conceptofmind/LLongMA-2-13b](https://huggingface.co/conceptofmind/LLongMA-2-13b) intended for summarization and literary analysis of fiction stories. It contains custom modeling code to use Flash Attention 2 during inference, which provides a significant speedup, especially at longer context lengths. To enable, pass `trust_remote_code=True,use_flash_attention=True` to `AutoModelForCausalLM`. ## How to Use ``` SOME_FICTION ### SUMMARY: ``` or ``` SOME_FICTION ### ANALYSIS: ``` ## Example outputs Outputs on the text of [_Waystation City_](https://www.uncannymagazine.com/article/waystation-city/) (6,904 tokens) ### Summary `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > A nineteen-seventies pair of twins approaches Micheala and Gerty at a cafe. They ask them to write down their story of their disappearance. Gerty takes notes on their story as they tell it. The twins' names are Daphne and Claude. The twins explain that they came from London, England and were leading a movement against fascism there. But in Waystation City, they were stuck in the limbo between timelines with no real way of returning home to continue their work. In fact, it was harder for them to remember their lives at home as time went on. They tell Gerty that they cannot stay and Micheala encourages them not to go down to the stairs where Waystation City's ferry drivers will take them home. Daphne writes an inscription on a wall, leaving her mark as one of the disappeared, but she and Claude relent. Together they approach the docks with Gerty and Micheala trailing behind. There, the twins discuss which hole to go through and they agree it is the one showing their hometown in the 1970s. Daphne goes first and then Claude after he slips on a wet surface. The ferry drivers see him but Gerty yells for him to return from the hole as she writes his story down with her notes. But he cannot hear her over the slip of metal slats turning into stone and wood on the docks. He jumps in anyway and disappears into the other world through the hole. Micheala leaves a cigarette and Gerty drops her ticket back home to show that she's decided to stay in Waystation City where it is possible to grow, change, and live in the spaces between. ### Analysis `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > Notes In this final chapter, there is a brief but important flashback to the twins' home in England at the beginning of the 1970s, when racism and fascism are rampant. It is interesting to note the author's choice to set the story at such a pivotal time for immigrants in Britain. The twins feel like they have to choose between being their "true" selves in their home country or staying where they are as Waystation citizens. But this was also the time when racism and fascism were so openly practiced that anti-racists were fighting against these institutions with all of their might. It is interesting how there's an implication here that it's only at the risk of their "true" identities that the twins can go back to England. In this chapter, Gerty changes her name as well, choosing the gender neutral Michael. She does not specify why she decided to change her name or what inspired her choice. The story implies, but never says outright, that she changed her name in part because of the twins' disappearance. This suggests an important question: how far must a person be willing to go in order to stay true to themselves? It is clear by this point that Gerty has changed quite a bit since coming to Waystation City and her decision to change her name reflects the extent to which she has altered her identity. She realizes that she has settled into a comfortable, new life here with new clothes, new hobbies, and friends. Ultimately, this chapter suggests that while it is important to stay true to one's past, there are also new things to discover and experience in the present. Gerty has done just this by settling in Waystation City, embracing her new life and even changing her name and identity with it. ## Training The model was trained on [emozilla/booksum-summary-analysis_llama-8192](https://huggingface.co/datasets/emozilla/booksum-summary-analysis_llama-8192), which is adapted from [kmfoda/booksum](https://huggingface.co/datasets/kmfoda/booksum). The training run was performed using [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl). The run can be viewed on [wandb](https://wandb.ai/emozilla/booksum/runs/53su11ol).

null

Non_BioNLP

GGML version of [emozilla/LLongMA-2-13b-storysummarizer](https://huggingface.co/emozilla/LLongMA-2-13b-storysummarizer). Use `n_ctx=8192` and `rope_freq_scale=0.5`. For an excellent user interface, check out [LM Studio](https://lmstudio.ai/). # LLongMA-2-13b-storysummarizer [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl) This is a fine-tuned version of [conceptofmind/LLongMA-2-13b](https://huggingface.co/conceptofmind/LLongMA-2-13b) intended for summarization and literary analysis of fiction stories. It contains custom modeling code to use Flash Attention 2 during inference, which provides a significant speedup, especially at longer context lengths. To enable, pass `trust_remote_code=True,use_flash_attention=True` to `AutoModelForCausalLM`. ## How to Use ``` SOME_FICTION ### SUMMARY: ``` or ``` SOME_FICTION ### ANALYSIS: ``` ## Example outputs Outputs on the text of [_Waystation City_](https://www.uncannymagazine.com/article/waystation-city/) (6,904 tokens) ### Summary `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > A nineteen-seventies pair of twins approaches Micheala and Gerty at a cafe. They ask them to write down their story of their disappearance. Gerty takes notes on their story as they tell it. The twins' names are Daphne and Claude. The twins explain that they came from London, England and were leading a movement against fascism there. But in Waystation City, they were stuck in the limbo between timelines with no real way of returning home to continue their work. In fact, it was harder for them to remember their lives at home as time went on. They tell Gerty that they cannot stay and Micheala encourages them not to go down to the stairs where Waystation City's ferry drivers will take them home. Daphne writes an inscription on a wall, leaving her mark as one of the disappeared, but she and Claude relent. Together they approach the docks with Gerty and Micheala trailing behind. There, the twins discuss which hole to go through and they agree it is the one showing their hometown in the 1970s. Daphne goes first and then Claude after he slips on a wet surface. The ferry drivers see him but Gerty yells for him to return from the hole as she writes his story down with her notes. But he cannot hear her over the slip of metal slats turning into stone and wood on the docks. He jumps in anyway and disappears into the other world through the hole. Micheala leaves a cigarette and Gerty drops her ticket back home to show that she's decided to stay in Waystation City where it is possible to grow, change, and live in the spaces between. ### Analysis `temperature=0.8`, `repetition_penalty=1.1`,`do_sample=True` > Notes In this final chapter, there is a brief but important flashback to the twins' home in England at the beginning of the 1970s, when racism and fascism are rampant. It is interesting to note the author's choice to set the story at such a pivotal time for immigrants in Britain. The twins feel like they have to choose between being their "true" selves in their home country or staying where they are as Waystation citizens. But this was also the time when racism and fascism were so openly practiced that anti-racists were fighting against these institutions with all of their might. It is interesting how there's an implication here that it's only at the risk of their "true" identities that the twins can go back to England. In this chapter, Gerty changes her name as well, choosing the gender neutral Michael. She does not specify why she decided to change her name or what inspired her choice. The story implies, but never says outright, that she changed her name in part because of the twins' disappearance. This suggests an important question: how far must a person be willing to go in order to stay true to themselves? It is clear by this point that Gerty has changed quite a bit since coming to Waystation City and her decision to change her name reflects the extent to which she has altered her identity. She realizes that she has settled into a comfortable, new life here with new clothes, new hobbies, and friends. Ultimately, this chapter suggests that while it is important to stay true to one's past, there are also new things to discover and experience in the present. Gerty has done just this by settling in Waystation City, embracing her new life and even changing her name and identity with it. ## Training The model was trained on [emozilla/booksum-summary-analysis_llama-8192](https://huggingface.co/datasets/emozilla/booksum-summary-analysis_llama-8192), which is adapted from [kmfoda/booksum](https://huggingface.co/datasets/kmfoda/booksum). The training run was performed using [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl). The run can be viewed on [wandb](https://wandb.ai/emozilla/booksum/runs/53su11ol).

{"datasets": ["emozilla/booksum-summary-analysis_llama-8192", "kmfoda/booksum"], "license": "apache-2.0"}

task

[ "SUMMARIZATION" ]

44,554

ruanchaves/bert-base-portuguese-cased-assin-entailment

ruanchaves

text-classification

[ "transformers", "pytorch", "bert", "text-classification", "pt", "dataset:assin", "autotrain_compatible", "region:us" ]

2023-03-27T18:09:12Z

2023-03-29T18:05:31+00:00

33

0

--- datasets: - assin language: pt inference: false --- # BERTimbau base for Recognizing Textual Entailment This is the [neuralmind/bert-base-portuguese-cased](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model finetuned for Recognizing Textual Entailment with the [ASSIN](https://huggingface.co/datasets/assin) dataset. This model is suitable for Portuguese. - Git Repo: [Evaluation of Portuguese Language Models](https://github.com/ruanchaves/eplm). - Demo: [Portuguese Textual Entailment](https://ruanchaves-portuguese-textual-entailment.hf.space) ### **Labels**: * 0 : There is no entailment between premise and hypothesis. * 1 : There is entailment between premise and hypothesis. * 2 : The premise is a paraphrase of the hypothesis. ## Full classification example ```python from transformers import AutoModelForSequenceClassification, AutoTokenizer, AutoConfig import numpy as np import torch from scipy.special import softmax model_name = "ruanchaves/bert-base-portuguese-cased-assin-entailment" s1 = "Os homens estão cuidadosamente colocando as malas no porta-malas de um carro." s2 = "Os homens estão colocando bagagens dentro do porta-malas de um carro." model = AutoModelForSequenceClassification.from_pretrained(model_name) tokenizer = AutoTokenizer.from_pretrained(model_name) config = AutoConfig.from_pretrained(model_name) model_input = tokenizer(*([s1], [s2]), padding=True, return_tensors="pt") with torch.no_grad(): output = model(**model_input) scores = output[0][0].detach().numpy() scores = softmax(scores) ranking = np.argsort(scores) ranking = ranking[::-1] for i in range(scores.shape[0]): l = config.id2label[ranking[i]] s = scores[ranking[i]] print(f"{i+1}) Label: {l} Score: {np.round(float(s), 4)}") ``` ## Citation Our research is ongoing, and we are currently working on describing our experiments in a paper, which will be published soon. In the meanwhile, if you would like to cite our work or models before the publication of the paper, please cite our [GitHub repository](https://github.com/ruanchaves/eplm): ``` @software{Chaves_Rodrigues_eplm_2023, author = {Chaves Rodrigues, Ruan and Tanti, Marc and Agerri, Rodrigo}, doi = {10.5281/zenodo.7781848}, month = {3}, title = {{Evaluation of Portuguese Language Models}}, url = {https://github.com/ruanchaves/eplm}, version = {1.0.0}, year = {2023} } ```

null

Non_BioNLP

# BERTimbau base for Recognizing Textual Entailment This is the [neuralmind/bert-base-portuguese-cased](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model finetuned for Recognizing Textual Entailment with the [ASSIN](https://huggingface.co/datasets/assin) dataset. This model is suitable for Portuguese. - Git Repo: [Evaluation of Portuguese Language Models](https://github.com/ruanchaves/eplm). - Demo: [Portuguese Textual Entailment](https://ruanchaves-portuguese-textual-entailment.hf.space) ### **Labels**: * 0 : There is no entailment between premise and hypothesis. * 1 : There is entailment between premise and hypothesis. * 2 : The premise is a paraphrase of the hypothesis. ## Full classification example ```python from transformers import AutoModelForSequenceClassification, AutoTokenizer, AutoConfig import numpy as np import torch from scipy.special import softmax model_name = "ruanchaves/bert-base-portuguese-cased-assin-entailment" s1 = "Os homens estão cuidadosamente colocando as malas no porta-malas de um carro." s2 = "Os homens estão colocando bagagens dentro do porta-malas de um carro." model = AutoModelForSequenceClassification.from_pretrained(model_name) tokenizer = AutoTokenizer.from_pretrained(model_name) config = AutoConfig.from_pretrained(model_name) model_input = tokenizer(*([s1], [s2]), padding=True, return_tensors="pt") with torch.no_grad(): output = model(**model_input) scores = output[0][0].detach().numpy() scores = softmax(scores) ranking = np.argsort(scores) ranking = ranking[::-1] for i in range(scores.shape[0]): l = config.id2label[ranking[i]] s = scores[ranking[i]] print(f"{i+1}) Label: {l} Score: {np.round(float(s), 4)}") ``` ## Citation Our research is ongoing, and we are currently working on describing our experiments in a paper, which will be published soon. In the meanwhile, if you would like to cite our work or models before the publication of the paper, please cite our [GitHub repository](https://github.com/ruanchaves/eplm): ``` @software{Chaves_Rodrigues_eplm_2023, author = {Chaves Rodrigues, Ruan and Tanti, Marc and Agerri, Rodrigo}, doi = {10.5281/zenodo.7781848}, month = {3}, title = {{Evaluation of Portuguese Language Models}}, url = {https://github.com/ruanchaves/eplm}, version = {1.0.0}, year = {2023} } ```

{"datasets": ["assin"], "language": "pt", "inference": false}

task

[ "TEXTUAL_ENTAILMENT" ]

44,555

neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic

neuralmagic

text-generation

[ "transformers", "safetensors", "qwen2", "text-generation", "deepseek", "fp8", "vllm", "conversational", "base_model:deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "base_model:quantized:deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "license:mit", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "compressed-tensors", "region:us" ]

2025-02-01T18:03:58Z

2025-02-27T08:10:20+00:00

1,069

1

--- base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B library_name: transformers license: mit tags: - deepseek - fp8 - vllm --- # DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic ## Model Overview - **Model Architecture:** Qwen2ForCausalLM - **Input:** Text - **Output:** Text - **Model Optimizations:** - **Weight quantization:** FP8 - **Activation quantization:** FP8 - **Release Date:** 2/5/2025 - **Version:** 1.0 - **Model Developers:** Neural Magic Quantized version of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B). ### Model Optimizations This model was obtained by quantizing the weights and activations of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B) to FP8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks are quantized. Weights are quantized using a symmetric per-channel scheme, whereas quantizations are quantized using a symmetric per-token scheme. [LLM Compressor](https://github.com/vllm-project/llm-compressor) is used for quantization. ## Use with vLLM This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams number_gpus = 1 model_name = "neuralmagic/DeepSeek-R1-Distill-Qwen-7B-dynamic" tokenizer = AutoTokenizer.from_pretrained(model_name) sampling_params = SamplingParams(temperature=0.6, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) llm = LLM(model=model_name, tensor_parallel_size=number_gpus, trust_remote_code=True) messages_list = [ [{"role": "user", "content": "Who are you? Please respond in pirate speak!"}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details. ## Creation This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below. ```python from transformers import AutoModelForCausalLM, AutoTokenizer from llmcompressor.modifiers.quantization import QuantizationModifier from llmcompressor.transformers import oneshot import os # Load model model_stub = "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B" model_name = model_stub.split("/")[-1] model = AutoModelForCausalLM.from_pretrained( model_stub, torch_dtype="auto", ) tokenizer = AutoTokenizer.from_pretrained(model_stub) # Configure the quantization algorithm and scheme recipe = QuantizationModifier( targets="Linear", scheme="FP8_DYNAMIC", ignore=["lm_head"], ) # Apply quantization oneshot( model=model, recipe=recipe, ) # Save to disk in compressed-tensors format save_path = model_name + "-FP8-dynamic model.save_pretrained(save_path) tokenizer.save_pretrained(save_path) print(f"Model and tokenizer saved to: {save_path}") ``` ## Evaluation The model was evaluated on OpenLLM Leaderboard [V1](https://huggingface.co/spaces/open-llm-leaderboard-old/open_llm_leaderboard) and [V2](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/), using the following commands: OpenLLM Leaderboard V1: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --tasks openllm \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` OpenLLM Leaderboard V2: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --apply_chat_template \ --fewshot_as_multiturn \ --tasks leaderboard \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` ### Accuracy <table> <thead> <tr> <th>Category</th> <th>Metric</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <th>Recovery</th> </tr> </thead> <tbody> <tr> <td rowspan="4">Reasoning</td> <td>AIME 2024 (pass@1)</td> <td>53.17</td> <td>53.17</td> <td>100%</td> </tr> <tr> <td>MATH-500 (pass@1)</td> <td>93.66</td> <td>93.62</td> <td>99.96%</td> </tr> <tr> <td>GPQA Diamond (pass@1)</td> <td>50.53</td> <td>50.28</td> <td>99.51%</td> </tr> <tr> <td>Average Score</td> <td>65.79</td> <td>65.69</td> <td>99.85%</td> </tr> <tr> <td rowspan="7">OpenLLM V1</td> <td>ARC-Challenge (Acc-Norm, 25-shot)</td> <td>50.51</td> <td>50.51</td> <td>100.0%</td> </tr> <tr> <td>GSM8K (Strict-Match, 5-shot)</td> <td>78.62</td> <td>79.83</td> <td>101.5%</td> </tr> <tr> <td>HellaSwag (Acc-Norm, 10-shot)</td> <td>61.90</td> <td>61.62</td> <td>99.6%</td> </tr> <tr> <td>MMLU (Acc, 5-shot)</td> <td>54.19</td> <td>53.76</td> <td>99.2%</td> </tr> <tr> <td>TruthfulQA (MC2, 0-shot)</td> <td>45.55</td> <td>46.14</td> <td>101.3%</td> </tr> <tr> <td>Winogrande (Acc, 5-shot)</td> <td>61.56</td> <td>60.54</td> <td>98.3%</td> </tr> <tr> <td>Average Score</td> <td>58.72</td> <td>58.73</td> <td>100.0%</td> </tr> <tr> <td rowspan="7">OpenLLM V2</td> <td>IFEval (Inst Level Strict Acc, 0-shot)</td> <td>39.38</td> <td>39.01</td> <td>99.1%</td> </tr> <tr> <td>BBH (Acc-Norm, 3-shot)</td> <td>6.97</td> <td>6.19</td> <td>---</td> </tr> <tr> <td>Math-Hard (Exact-Match, 4-shot)</td> <td>0.00</td> <td>0.00</td> <td>---</td> </tr> <tr> <td>GPQA (Acc-Norm, 0-shot)</td> <td>1.81</td> <td>1.63</td> <td>---</td> </tr> <tr> <td>MUSR (Acc-Norm, 0-shot)</td> <td>4.68</td> <td>5.08</td> <td>---</td> </tr> <tr> <td>MMLU-Pro (Acc, 5-shot)</td> <td>1.66</td> <td>1.76</td> <td>---</td> </tr> <tr> <td>Average Score</td> <td>9.08</td> <td>8.94</td> <td>---</td> </tr> <tr> <td rowspan="4">Coding</td> <td>HumanEval (pass@1)</td> <td>40.80</td> <td>39.50</td> <td>96.8%</td> </tr> <tr> <td>HumanEval (pass@10)</td> <td>64.40</td> <td>62.10</td> <td>96.4%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>38.50</td> <td>37.20</td> <td>96.6%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>60.40</td> <td>59.30</td> <td>98.2%</td> </tr> </tbody> </table> ## Inference Performance This model achieves up to 1.4x speedup in single-stream deployment and up to 1.2x speedup in multi-stream asynchronous deployment, depending on hardware and use-case scenario. The following performance benchmarks were conducted with [vLLM](https://docs.vllm.ai/en/latest/) version 0.7.2, and [GuideLLM](https://github.com/neuralmagic/guidellm). <details> <summary>Benchmarking Command</summary> ``` guidellm --model neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic --target "http://localhost:8000/v1" --data-type emulated --data "prompt_tokens=<prompt_tokens>,generated_tokens=<generated_tokens>" --max seconds 360 --backend aiohttp_server ``` </details> ### Single-stream performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following 256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat 512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation 768 / 128</th> <th style="text-align: center;" colspan="2" >RAG 1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion 256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing 1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization 4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG 10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>2.9</td> <td>1576</td> <td>5.7</td> <td>788</td> <td>2.9</td> <td>1535</td> <td>3.0</td> <td>1496</td> <td>22.6</td> <td>199</td> <td>23.2</td> <td>194</td> <td>12.1</td> <td>370</td> <td>38.5</td> <td>117</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.56</td> <td>1.8</td> <td>2495</td> <td>3.7</td> <td>1223</td> <td>1.9</td> <td>2384</td> <td>1.9</td> <td>2393</td> <td>14.3</td> <td>315</td> <td>14.8</td> <td>304</td> <td>7.9</td> <td>572</td> <td>25.3</td> <td>178</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>2.41</td> <td>1.1</td> <td>4086</td> <td>2.3</td> <td>1998</td> <td>1.2</td> <td>3783</td> <td>1.3</td> <td>3527</td> <td>8.6</td> <td>526</td> <td>8.8</td> <td>512</td> <td>5.2</td> <td>860</td> <td>22.7</td> <td>198</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>1.4</td> <td>1389</td> <td>2.9</td> <td>691</td> <td>1.5</td> <td>1358</td> <td>1.5</td> <td>1329</td> <td>11.5</td> <td>175</td> <td>11.6</td> <td>174</td> <td>6.2</td> <td>326</td> <td>21.5</td> <td>93</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.28</td> <td>1.1</td> <td>1850</td> <td>2.2</td> <td>905</td> <td>1.1</td> <td>1807</td> <td>1.1</td> <td>1750</td> <td>8.6</td> <td>233</td> <td>8.7</td> <td>230</td> <td>4.7</td> <td>431</td> <td>23.1</td> <td>87</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.72</td> <td>0.8</td> <td>2575</td> <td>1.5</td> <td>1298</td> <td>0.8</td> <td>2461</td> <td>0.8</td> <td>2382</td> <td>6.1</td> <td>331</td> <td>6.2</td> <td>323</td> <td>3.6</td> <td>566</td> <td>22.7</td> <td>89</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>0.9</td> <td>1161</td> <td>1.9</td> <td>579</td> <td>1.0</td> <td>1138</td> <td>1.0</td> <td>1121</td> <td>7.5</td> <td>146</td> <td>7.6</td> <td>145</td> <td>3.9</td> <td>279</td> <td>15.4</td> <td>71</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.34</td> <td>0.7</td> <td>1585</td> <td>1.4</td> <td>786</td> <td>0.7</td> <td>1577</td> <td>0.7</td> <td>1524</td> <td>5.3</td> <td>207</td> <td>5.5</td> <td>197</td> <td>2.9</td> <td>382</td> <td>14.3</td> <td>77</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.33</td> <td>0.7</td> <td>1590</td> <td>1.4</td> <td>793</td> <td>0.7</td> <td>1549</td> <td>0.7</td> <td>1509</td> <td>5.4</td> <td>201</td> <td>5.5</td> <td>198</td> <td>2.9</td> <td>381</td> <td>14.0</td> <td>78</td> </tr> </tbody> </table> **Use case profiles: prompt tokens / generation tokens **QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025). ### Multi-stream asynchronous performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following 256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat 512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation 768 / 128</th> <th style="text-align: center;" colspan="2" >RAG 1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion 256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing 1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization 4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG 10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>14.9</td> <td>67138</td> <td>7.1</td> <td>32094</td> <td>7.4</td> <td>33096</td> <td>5.9</td> <td>26480</td> <td>2.0</td> <td>9004</td> <td>1.5</td> <td>6639</td> <td>1.1</td> <td>4938</td> <td>0.3</td> <td>1151</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.36</td> <td>20.2</td> <td>90956</td> <td>8.8</td> <td>39786</td> <td>10.2</td> <td>45963</td> <td>8.1</td> <td>36596</td> <td>3.1</td> <td>13968</td> <td>2.1</td> <td>9629</td> <td>1.4</td> <td>6374</td> <td>0.3</td> <td>1429</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.00</td> <td>13.3</td> <td>59681</td> <td>6.1</td> <td>27633</td> <td>5.9</td> <td>26689</td> <td>4.7</td> <td>20944</td> <td>2.9</td> <td>13108</td> <td>1.9</td> <td>8355</td> <td>1.0</td> <td>4362</td> <td>0.3</td> <td>1170</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>26.4</td> <td>53073</td> <td>13.0</td> <td>26213</td> <td>14.5</td> <td>29110</td> <td>11.4</td> <td>22936</td> <td>4.4</td> <td>8749</td> <td>3.3</td> <td>6680</td> <td>2.3</td> <td>4634</td> <td>0.5</td> <td>1105</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.27</td> <td>34.3</td> <td>69009</td> <td>14.8</td> <td>29791</td> <td>19.0</td> <td>38214</td> <td>15.7</td> <td>31598</td> <td>5.6</td> <td>11186</td> <td>4.2</td> <td>8350</td> <td>3.0</td> <td>6020</td> <td>0.7</td> <td>1328</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>0.93</td> <td>23.9</td> <td>47993</td> <td>12.0</td> <td>24194</td> <td>12.5</td> <td>25239</td> <td>10.0</td> <td>20029</td> <td>4.5</td> <td>9055</td> <td>3.3</td> <td>6681</td> <td>2.1</td> <td>4156</td> <td>0.5</td> <td>1043</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>54.3</td> <td>59410</td> <td>26.0</td> <td>28440</td> <td>32.1</td> <td>35154</td> <td>26.7</td> <td>29190</td> <td>8.0</td> <td>8700</td> <td>6.6</td> <td>7275</td> <td>5.2</td> <td>5669</td> <td>1.2</td> <td>1266</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.16</td> <td>62.9</td> <td>68818</td> <td>30.3</td> <td>33196</td> <td>39.4</td> <td>43132</td> <td>31.1</td> <td>34073</td> <td>9.2</td> <td>10058</td> <td>7.1</td> <td>7748</td> <td>6.1</td> <td>6714</td> <td>1.3</td> <td>1415</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.02</td> <td>56.2</td> <td>61483</td> <td>26.7</td> <td>29243</td> <td>32.5</td> <td>35592</td> <td>26.9</td> <td>29461</td> <td>8.3</td> <td>9072</td> <td>6.4</td> <td>7027</td> <td>5.2</td> <td>5731</td> <td>1.2</td> <td>1291</td> </tr> </tbody> </table> **Use case profiles: prompt tokens / generation tokens **QPS: Queries per second. **QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025).

null

Non_BioNLP

# DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic ## Model Overview - **Model Architecture:** Qwen2ForCausalLM - **Input:** Text - **Output:** Text - **Model Optimizations:** - **Weight quantization:** FP8 - **Activation quantization:** FP8 - **Release Date:** 2/5/2025 - **Version:** 1.0 - **Model Developers:** Neural Magic Quantized version of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B). ### Model Optimizations This model was obtained by quantizing the weights and activations of [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B) to FP8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks are quantized. Weights are quantized using a symmetric per-channel scheme, whereas quantizations are quantized using a symmetric per-token scheme. [LLM Compressor](https://github.com/vllm-project/llm-compressor) is used for quantization. ## Use with vLLM This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams number_gpus = 1 model_name = "neuralmagic/DeepSeek-R1-Distill-Qwen-7B-dynamic" tokenizer = AutoTokenizer.from_pretrained(model_name) sampling_params = SamplingParams(temperature=0.6, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) llm = LLM(model=model_name, tensor_parallel_size=number_gpus, trust_remote_code=True) messages_list = [ [{"role": "user", "content": "Who are you? Please respond in pirate speak!"}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details. ## Creation This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below. ```python from transformers import AutoModelForCausalLM, AutoTokenizer from llmcompressor.modifiers.quantization import QuantizationModifier from llmcompressor.transformers import oneshot import os # Load model model_stub = "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B" model_name = model_stub.split("/")[-1] model = AutoModelForCausalLM.from_pretrained( model_stub, torch_dtype="auto", ) tokenizer = AutoTokenizer.from_pretrained(model_stub) # Configure the quantization algorithm and scheme recipe = QuantizationModifier( targets="Linear", scheme="FP8_DYNAMIC", ignore=["lm_head"], ) # Apply quantization oneshot( model=model, recipe=recipe, ) # Save to disk in compressed-tensors format save_path = model_name + "-FP8-dynamic model.save_pretrained(save_path) tokenizer.save_pretrained(save_path) print(f"Model and tokenizer saved to: {save_path}") ``` ## Evaluation The model was evaluated on OpenLLM Leaderboard [V1](https://huggingface.co/spaces/open-llm-leaderboard-old/open_llm_leaderboard) and [V2](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/), using the following commands: OpenLLM Leaderboard V1: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --tasks openllm \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` OpenLLM Leaderboard V2: ``` lm_eval \ --model vllm \ --model_args pretrained="neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ --apply_chat_template \ --fewshot_as_multiturn \ --tasks leaderboard \ --write_out \ --batch_size auto \ --output_path output_dir \ --show_config ``` ### Accuracy <table> <thead> <tr> <th>Category</th> <th>Metric</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <th>Recovery</th> </tr> </thead> <tbody> <tr> <td rowspan="4">Reasoning</td> <td>AIME 2024 (pass@1)</td> <td>53.17</td> <td>53.17</td> <td>100%</td> </tr> <tr> <td>MATH-500 (pass@1)</td> <td>93.66</td> <td>93.62</td> <td>99.96%</td> </tr> <tr> <td>GPQA Diamond (pass@1)</td> <td>50.53</td> <td>50.28</td> <td>99.51%</td> </tr> <tr> <td>Average Score</td> <td>65.79</td> <td>65.69</td> <td>99.85%</td> </tr> <tr> <td rowspan="7">OpenLLM V1</td> <td>ARC-Challenge (Acc-Norm, 25-shot)</td> <td>50.51</td> <td>50.51</td> <td>100.0%</td> </tr> <tr> <td>GSM8K (Strict-Match, 5-shot)</td> <td>78.62</td> <td>79.83</td> <td>101.5%</td> </tr> <tr> <td>HellaSwag (Acc-Norm, 10-shot)</td> <td>61.90</td> <td>61.62</td> <td>99.6%</td> </tr> <tr> <td>MMLU (Acc, 5-shot)</td> <td>54.19</td> <td>53.76</td> <td>99.2%</td> </tr> <tr> <td>TruthfulQA (MC2, 0-shot)</td> <td>45.55</td> <td>46.14</td> <td>101.3%</td> </tr> <tr> <td>Winogrande (Acc, 5-shot)</td> <td>61.56</td> <td>60.54</td> <td>98.3%</td> </tr> <tr> <td>Average Score</td> <td>58.72</td> <td>58.73</td> <td>100.0%</td> </tr> <tr> <td rowspan="7">OpenLLM V2</td> <td>IFEval (Inst Level Strict Acc, 0-shot)</td> <td>39.38</td> <td>39.01</td> <td>99.1%</td> </tr> <tr> <td>BBH (Acc-Norm, 3-shot)</td> <td>6.97</td> <td>6.19</td> <td>---</td> </tr> <tr> <td>Math-Hard (Exact-Match, 4-shot)</td> <td>0.00</td> <td>0.00</td> <td>---</td> </tr> <tr> <td>GPQA (Acc-Norm, 0-shot)</td> <td>1.81</td> <td>1.63</td> <td>---</td> </tr> <tr> <td>MUSR (Acc-Norm, 0-shot)</td> <td>4.68</td> <td>5.08</td> <td>---</td> </tr> <tr> <td>MMLU-Pro (Acc, 5-shot)</td> <td>1.66</td> <td>1.76</td> <td>---</td> </tr> <tr> <td>Average Score</td> <td>9.08</td> <td>8.94</td> <td>---</td> </tr> <tr> <td rowspan="4">Coding</td> <td>HumanEval (pass@1)</td> <td>40.80</td> <td>39.50</td> <td>96.8%</td> </tr> <tr> <td>HumanEval (pass@10)</td> <td>64.40</td> <td>62.10</td> <td>96.4%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>38.50</td> <td>37.20</td> <td>96.6%</td> </tr> <tr> <td>HumanEval+ (pass@10)</td> <td>60.40</td> <td>59.30</td> <td>98.2%</td> </tr> </tbody> </table> ## Inference Performance This model achieves up to 1.4x speedup in single-stream deployment and up to 1.2x speedup in multi-stream asynchronous deployment, depending on hardware and use-case scenario. The following performance benchmarks were conducted with [vLLM](https://docs.vllm.ai/en/latest/) version 0.7.2, and [GuideLLM](https://github.com/neuralmagic/guidellm). <details> <summary>Benchmarking Command</summary> ``` guidellm --model neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic --target "http://localhost:8000/v1" --data-type emulated --data "prompt_tokens=<prompt_tokens>,generated_tokens=<generated_tokens>" --max seconds 360 --backend aiohttp_server ``` </details> ### Single-stream performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following 256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat 512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation 768 / 128</th> <th style="text-align: center;" colspan="2" >RAG 1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion 256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing 1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization 4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG 10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> <th>Latency (s)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>2.9</td> <td>1576</td> <td>5.7</td> <td>788</td> <td>2.9</td> <td>1535</td> <td>3.0</td> <td>1496</td> <td>22.6</td> <td>199</td> <td>23.2</td> <td>194</td> <td>12.1</td> <td>370</td> <td>38.5</td> <td>117</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.56</td> <td>1.8</td> <td>2495</td> <td>3.7</td> <td>1223</td> <td>1.9</td> <td>2384</td> <td>1.9</td> <td>2393</td> <td>14.3</td> <td>315</td> <td>14.8</td> <td>304</td> <td>7.9</td> <td>572</td> <td>25.3</td> <td>178</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>2.41</td> <td>1.1</td> <td>4086</td> <td>2.3</td> <td>1998</td> <td>1.2</td> <td>3783</td> <td>1.3</td> <td>3527</td> <td>8.6</td> <td>526</td> <td>8.8</td> <td>512</td> <td>5.2</td> <td>860</td> <td>22.7</td> <td>198</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>1.4</td> <td>1389</td> <td>2.9</td> <td>691</td> <td>1.5</td> <td>1358</td> <td>1.5</td> <td>1329</td> <td>11.5</td> <td>175</td> <td>11.6</td> <td>174</td> <td>6.2</td> <td>326</td> <td>21.5</td> <td>93</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.28</td> <td>1.1</td> <td>1850</td> <td>2.2</td> <td>905</td> <td>1.1</td> <td>1807</td> <td>1.1</td> <td>1750</td> <td>8.6</td> <td>233</td> <td>8.7</td> <td>230</td> <td>4.7</td> <td>431</td> <td>23.1</td> <td>87</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.72</td> <td>0.8</td> <td>2575</td> <td>1.5</td> <td>1298</td> <td>0.8</td> <td>2461</td> <td>0.8</td> <td>2382</td> <td>6.1</td> <td>331</td> <td>6.2</td> <td>323</td> <td>3.6</td> <td>566</td> <td>22.7</td> <td>89</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>0.9</td> <td>1161</td> <td>1.9</td> <td>579</td> <td>1.0</td> <td>1138</td> <td>1.0</td> <td>1121</td> <td>7.5</td> <td>146</td> <td>7.6</td> <td>145</td> <td>3.9</td> <td>279</td> <td>15.4</td> <td>71</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.34</td> <td>0.7</td> <td>1585</td> <td>1.4</td> <td>786</td> <td>0.7</td> <td>1577</td> <td>0.7</td> <td>1524</td> <td>5.3</td> <td>207</td> <td>5.5</td> <td>197</td> <td>2.9</td> <td>382</td> <td>14.3</td> <td>77</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.33</td> <td>0.7</td> <td>1590</td> <td>1.4</td> <td>793</td> <td>0.7</td> <td>1549</td> <td>0.7</td> <td>1509</td> <td>5.4</td> <td>201</td> <td>5.5</td> <td>198</td> <td>2.9</td> <td>381</td> <td>14.0</td> <td>78</td> </tr> </tbody> </table> **Use case profiles: prompt tokens / generation tokens **QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025). ### Multi-stream asynchronous performance (measured with vLLM version 0.7.2) <table> <thead> <tr> <th></th> <th></th> <th></th> <th style="text-align: center;" colspan="2" >Instruction Following 256 / 128</th> <th style="text-align: center;" colspan="2" >Multi-turn Chat 512 / 256</th> <th style="text-align: center;" colspan="2" >Docstring Generation 768 / 128</th> <th style="text-align: center;" colspan="2" >RAG 1024 / 128</th> <th style="text-align: center;" colspan="2" >Code Completion 256 / 1024</th> <th style="text-align: center;" colspan="2" >Code Fixing 1024 / 1024</th> <th style="text-align: center;" colspan="2" >Large Summarization 4096 / 512</th> <th style="text-align: center;" colspan="2" >Large RAG 10240 / 1536</th> </tr> <tr> <th>Hardware</th> <th>Model</th> <th>Average cost reduction</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> <th>Maximum throughput (QPS)</th> <th>QPD</th> </tr> </thead> <tbody style="text-align: center" > <tr> <th rowspan="3" valign="top">A6000x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>14.9</td> <td>67138</td> <td>7.1</td> <td>32094</td> <td>7.4</td> <td>33096</td> <td>5.9</td> <td>26480</td> <td>2.0</td> <td>9004</td> <td>1.5</td> <td>6639</td> <td>1.1</td> <td>4938</td> <td>0.3</td> <td>1151</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.36</td> <td>20.2</td> <td>90956</td> <td>8.8</td> <td>39786</td> <td>10.2</td> <td>45963</td> <td>8.1</td> <td>36596</td> <td>3.1</td> <td>13968</td> <td>2.1</td> <td>9629</td> <td>1.4</td> <td>6374</td> <td>0.3</td> <td>1429</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.00</td> <td>13.3</td> <td>59681</td> <td>6.1</td> <td>27633</td> <td>5.9</td> <td>26689</td> <td>4.7</td> <td>20944</td> <td>2.9</td> <td>13108</td> <td>1.9</td> <td>8355</td> <td>1.0</td> <td>4362</td> <td>0.3</td> <td>1170</td> </tr> <tr> <th rowspan="3" valign="top">A100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>26.4</td> <td>53073</td> <td>13.0</td> <td>26213</td> <td>14.5</td> <td>29110</td> <td>11.4</td> <td>22936</td> <td>4.4</td> <td>8749</td> <td>3.3</td> <td>6680</td> <td>2.3</td> <td>4634</td> <td>0.5</td> <td>1105</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8</th> <td>1.27</td> <td>34.3</td> <td>69009</td> <td>14.8</td> <td>29791</td> <td>19.0</td> <td>38214</td> <td>15.7</td> <td>31598</td> <td>5.6</td> <td>11186</td> <td>4.2</td> <td>8350</td> <td>3.0</td> <td>6020</td> <td>0.7</td> <td>1328</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>0.93</td> <td>23.9</td> <td>47993</td> <td>12.0</td> <td>24194</td> <td>12.5</td> <td>25239</td> <td>10.0</td> <td>20029</td> <td>4.5</td> <td>9055</td> <td>3.3</td> <td>6681</td> <td>2.1</td> <td>4156</td> <td>0.5</td> <td>1043</td> </tr> <tr> <th rowspan="3" valign="top">H100x1</th> <th>deepseek-ai/DeepSeek-R1-Distill-Qwen-7B</th> <td>---</td> <td>54.3</td> <td>59410</td> <td>26.0</td> <td>28440</td> <td>32.1</td> <td>35154</td> <td>26.7</td> <td>29190</td> <td>8.0</td> <td>8700</td> <td>6.6</td> <td>7275</td> <td>5.2</td> <td>5669</td> <td>1.2</td> <td>1266</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic</th> <td>1.16</td> <td>62.9</td> <td>68818</td> <td>30.3</td> <td>33196</td> <td>39.4</td> <td>43132</td> <td>31.1</td> <td>34073</td> <td>9.2</td> <td>10058</td> <td>7.1</td> <td>7748</td> <td>6.1</td> <td>6714</td> <td>1.3</td> <td>1415</td> </tr> <tr> <th>neuralmagic/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16</th> <td>1.02</td> <td>56.2</td> <td>61483</td> <td>26.7</td> <td>29243</td> <td>32.5</td> <td>35592</td> <td>26.9</td> <td>29461</td> <td>8.3</td> <td>9072</td> <td>6.4</td> <td>7027</td> <td>5.2</td> <td>5731</td> <td>1.2</td> <td>1291</td> </tr> </tbody> </table> **Use case profiles: prompt tokens / generation tokens **QPS: Queries per second. **QPD: Queries per dollar, based on on-demand cost at [Lambda Labs](https://lambdalabs.com/service/gpu-cloud) (observed on 2/18/2025).

{"base_model": "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B", "library_name": "transformers", "license": "mit", "tags": ["deepseek", "fp8", "vllm"]}

task

[ "SUMMARIZATION" ]

44,556

AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-product-reviews

AdapterHub

null

[ "adapter-transformers", "xlm-roberta", "dataset:UKPLab/m2qa", "region:us" ]

2024-06-05T20:09:06Z

2024-12-11T11:28:56+00:00

8

0

--- datasets: - UKPLab/m2qa tags: - adapter-transformers - xlm-roberta --- # M2QA Adapter: Domain Adapter for MAD-X+Domain Setup This adapter is part of the M2QA publication to achieve language and domain transfer via adapters. 📃 Paper: [https://aclanthology.org/2024.findings-emnlp.365/](https://aclanthology.org/2024.findings-emnlp.365/) 🏗️ GitHub repo: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa) 💾 Hugging Face Dataset: [https://huggingface.co/UKPLab/m2qa](https://huggingface.co/UKPLab/m2qa) **Important:** This adapter only works together with the MAD-X language adapters and the M2QA QA head adapter. This [adapter](https://adapterhub.ml) for the `xlm-roberta-base` model that was trained using the **[Adapters](https://github.com/Adapter-Hub/adapters)** library. For detailed training details see our paper or GitHub repository: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa). You can find the evaluation results for this adapter on the M2QA dataset in the GitHub repo and in the paper. ## Usage First, install `adapters`: ``` pip install -U adapters ``` Now, the adapter can be loaded and activated like this: ```python from adapters import AutoAdapterModel from adapters.composition import Stack model = AutoAdapterModel.from_pretrained("xlm-roberta-base") # 1. Load language adapter language_adapter_name = model.load_adapter("de/wiki@ukp") # MAD-X+Domain uses the MAD-X language adapter # 2. Load domain adapter domain_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-product-reviews") # 3. Load QA head adapter qa_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-qa-head") # 4. Activate them via the adapter stack model.active_adapters = Stack(language_adapter_name, domain_adapter_name, qa_adapter_name) ``` See our repository for more information: See https://github.com/UKPLab/m2qa/tree/main/Experiments/mad-x-domain ## Contact Leon Engländer: - [HuggingFace Profile](https://huggingface.co/lenglaender) - [GitHub](https://github.com/lenglaender) - [Twitter](https://x.com/LeonEnglaender) ## Citation ``` @inproceedings{englander-etal-2024-m2qa, title = "M2QA: Multi-domain Multilingual Question Answering", author = {Engl{\"a}nder, Leon and Sterz, Hannah and Poth, Clifton A and Pfeiffer, Jonas and Kuznetsov, Ilia and Gurevych, Iryna}, booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024", month = nov, year = "2024", address = "Miami, Florida, USA", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.findings-emnlp.365", pages = "6283--6305", } ```

null

Non_BioNLP

# M2QA Adapter: Domain Adapter for MAD-X+Domain Setup This adapter is part of the M2QA publication to achieve language and domain transfer via adapters. 📃 Paper: [https://aclanthology.org/2024.findings-emnlp.365/](https://aclanthology.org/2024.findings-emnlp.365/) 🏗️ GitHub repo: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa) 💾 Hugging Face Dataset: [https://huggingface.co/UKPLab/m2qa](https://huggingface.co/UKPLab/m2qa) **Important:** This adapter only works together with the MAD-X language adapters and the M2QA QA head adapter. This [adapter](https://adapterhub.ml) for the `xlm-roberta-base` model that was trained using the **[Adapters](https://github.com/Adapter-Hub/adapters)** library. For detailed training details see our paper or GitHub repository: [https://github.com/UKPLab/m2qa](https://github.com/UKPLab/m2qa). You can find the evaluation results for this adapter on the M2QA dataset in the GitHub repo and in the paper. ## Usage First, install `adapters`: ``` pip install -U adapters ``` Now, the adapter can be loaded and activated like this: ```python from adapters import AutoAdapterModel from adapters.composition import Stack model = AutoAdapterModel.from_pretrained("xlm-roberta-base") # 1. Load language adapter language_adapter_name = model.load_adapter("de/wiki@ukp") # MAD-X+Domain uses the MAD-X language adapter # 2. Load domain adapter domain_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-product-reviews") # 3. Load QA head adapter qa_adapter_name = model.load_adapter("AdapterHub/m2qa-xlm-roberta-base-mad-x-domain-qa-head") # 4. Activate them via the adapter stack model.active_adapters = Stack(language_adapter_name, domain_adapter_name, qa_adapter_name) ``` See our repository for more information: See https://github.com/UKPLab/m2qa/tree/main/Experiments/mad-x-domain ## Contact Leon Engländer: - [HuggingFace Profile](https://huggingface.co/lenglaender) - [GitHub](https://github.com/lenglaender) - [Twitter](https://x.com/LeonEnglaender) ## Citation ``` @inproceedings{englander-etal-2024-m2qa, title = "M2QA: Multi-domain Multilingual Question Answering", author = {Engl{\"a}nder, Leon and Sterz, Hannah and Poth, Clifton A and Pfeiffer, Jonas and Kuznetsov, Ilia and Gurevych, Iryna}, booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024", month = nov, year = "2024", address = "Miami, Florida, USA", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.findings-emnlp.365", pages = "6283--6305", } ```

{"datasets": ["UKPLab/m2qa"], "tags": ["adapter-transformers", "xlm-roberta"]}

task

[ "QUESTION_ANSWERING" ]

44,557

HelgeKn/SemEval-multi-class-4

HelgeKn

text-classification

[ "setfit", "safetensors", "mpnet", "sentence-transformers", "text-classification", "generated_from_setfit_trainer", "arxiv:2209.11055", "base_model:sentence-transformers/paraphrase-mpnet-base-v2", "base_model:finetune:sentence-transformers/paraphrase-mpnet-base-v2", "model-index", "region:us" ]

2023-12-14T13:05:58Z

2023-12-14T13:07:17+00:00

49

0

--- base_model: sentence-transformers/paraphrase-mpnet-base-v2 library_name: setfit metrics: - accuracy pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 'The Alavas worked themselves to the bone in the last period , and English and San Emeterio ( 65-75 ) had already made it clear that they were not going to let anyone take away what they had earned during the first thirty minutes . ' - text: 'To break the uncomfortable silence , Haney began to talk . ' - text: 'For the treatment of non-small cell lung cancer , the effects of Alimta were compared with those of docetaxel ( another anticancer medicine ) in one study involving 571 patients with locally advanced or metastatic disease who had received chemotherapy in the past . ' - text: 'As we all know , a few minutes before the end of the game ( that their team had already won ) , both players deliberately wasted time which made the referee show the second yellow card to both of them . ' - text: 'In contrast , patients whose cancer was affecting squamous cells had shorter survival times if they received Alimta . ' inference: true model-index: - name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2 results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: accuracy value: 0.14172185430463577 name: Accuracy --- # SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - **Classification head:** a [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 7 classes    ### Model Sources - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit) - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels | Label | Examples | |:------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 | <ul><li>'Eventually little French farmers and their little French farmwives came out of their stone houses and put their hands above their tiny eyes and squinted at us . '</li><li>'Mr. Neuberger realized that , although of Italian ancestry , Mr. Mariotta still could qualify as a minority person since he was born in Puerto Rico . '</li><li>"Biggest trouble was scared family who could n't get a phone line through , and spent a really horrible hour not knowing . "</li></ul> | | 4 | <ul><li>'`` To ring for even one service at this tower , we have to scrape , `` says Mr. Hammond , a retired water-authority worker . `` '</li><li>"`` It 's my line of work `` , he said "</li><li>'One writer , signing his letter as `` Red-blooded , balanced male , `` remarked on the `` frequency of women fainting in peals , `` and suggested that they `` settle back into their traditional role of making tea at meetings . `` '</li></ul> | | 5 | <ul><li>'Of course On Thursday , Haney mailed the monthly check for separate maintenance to his wife Lolly , and wished the stranger could do something about her '</li><li>"On the Right , the tone was set by Jacques Chirac , who declared in 1976 that `` 900,000 unemployed would not become a problem in a country with 2 million of foreign workers , '' and on the Left by Michel Rocard explaining in 1990 that France `` can not accommodate all the world 's misery . '' "</li><li>"But the council 's program to attract and train ringers is only partly successful , says Mr. Baldwin . "</li></ul> | | 6 | <ul><li>'3 -RRB- Republican congressional representatives , because of their belief in a minimalist state , are less willing to engage in local benefit-seeking than are Democratic members of Congress . '</li><li>'As we know , voters tend to favor Republicans more in races for president than in those for Congress . '</li><li>'That is the way the system works . '</li></ul> | | 2 | <ul><li>'-- Students should move up the educational ladder as their academic potential allows . '</li><li>'The next day , Sunday , the hangover reminded Haney where he had been the night before . '</li><li>'-- In most states , increasing expenditures on education , in our current circumstances , will probably make things worse , not better . '</li></ul> | | 0 | <ul><li>'Then your focus will go to an input text box where you can type your function . '</li><li>"I might have got hit by that truck if it was n't for you . "</li><li>"Second , it explains why voters hold Congress in disdain but generally love their own congressional representatives : Any individual legislator 's constituents appreciate the specific benefits that the legislator wins for them but not the overall cost associated with every other legislator doing likewise for his own constituency . "</li></ul> | | 3 | <ul><li>"It was the most exercise we 'd had all morning and it was followed by our driving immediately to the nearest watering hole . "</li><li>'Alimta is used together with cisplatin ( another anticancer medicine ) when the cancer is unresectable ( cannot be removed by surgery alone ) and malignant ( has spread , or is likely to spread easily , to other parts of the body ) , in patients who have not received chemotherapy ( medicines for cancer ) before advanced or metastatic non-small cell lung cancer that is not affecting the squamous cells . '</li><li>'If it is , it will be treated as an operator , if it is not , it will be treated as a user function . '</li></ul> | ## Evaluation ### Metrics | Label | Accuracy | |:--------|:---------| | **all** | 0.1417 | ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("HelgeKn/SemEval-multi-class-4") # Run inference preds = model("To break the uncomfortable silence , Haney began to talk . ") ```     ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 4 | 27.1786 | 74 | | Label | Training Sample Count | |:------|:----------------------| | 0 | 4 | | 1 | 4 | | 2 | 4 | | 3 | 4 | | 4 | 4 | | 5 | 4 | | 6 | 4 | ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (2, 2) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:------:|:----:|:-------------:|:---------------:| | 0.0143 | 1 | 0.2446 | - | | 0.7143 | 50 | 0.0612 | - | | 1.4286 | 100 | 0.0078 | - | ### Framework Versions - Python: 3.9.13 - SetFit: 1.0.1 - Sentence Transformers: 2.2.2 - Transformers: 4.36.0 - PyTorch: 2.1.1+cpu - Datasets: 2.15.0 - Tokenizers: 0.15.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

# SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - **Classification head:** a [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 7 classes    ### Model Sources - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit) - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels | Label | Examples | |:------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 | <ul><li>'Eventually little French farmers and their little French farmwives came out of their stone houses and put their hands above their tiny eyes and squinted at us . '</li><li>'Mr. Neuberger realized that , although of Italian ancestry , Mr. Mariotta still could qualify as a minority person since he was born in Puerto Rico . '</li><li>"Biggest trouble was scared family who could n't get a phone line through , and spent a really horrible hour not knowing . "</li></ul> | | 4 | <ul><li>'`` To ring for even one service at this tower , we have to scrape , `` says Mr. Hammond , a retired water-authority worker . `` '</li><li>"`` It 's my line of work `` , he said "</li><li>'One writer , signing his letter as `` Red-blooded , balanced male , `` remarked on the `` frequency of women fainting in peals , `` and suggested that they `` settle back into their traditional role of making tea at meetings . `` '</li></ul> | | 5 | <ul><li>'Of course On Thursday , Haney mailed the monthly check for separate maintenance to his wife Lolly , and wished the stranger could do something about her '</li><li>"On the Right , the tone was set by Jacques Chirac , who declared in 1976 that `` 900,000 unemployed would not become a problem in a country with 2 million of foreign workers , '' and on the Left by Michel Rocard explaining in 1990 that France `` can not accommodate all the world 's misery . '' "</li><li>"But the council 's program to attract and train ringers is only partly successful , says Mr. Baldwin . "</li></ul> | | 6 | <ul><li>'3 -RRB- Republican congressional representatives , because of their belief in a minimalist state , are less willing to engage in local benefit-seeking than are Democratic members of Congress . '</li><li>'As we know , voters tend to favor Republicans more in races for president than in those for Congress . '</li><li>'That is the way the system works . '</li></ul> | | 2 | <ul><li>'-- Students should move up the educational ladder as their academic potential allows . '</li><li>'The next day , Sunday , the hangover reminded Haney where he had been the night before . '</li><li>'-- In most states , increasing expenditures on education , in our current circumstances , will probably make things worse , not better . '</li></ul> | | 0 | <ul><li>'Then your focus will go to an input text box where you can type your function . '</li><li>"I might have got hit by that truck if it was n't for you . "</li><li>"Second , it explains why voters hold Congress in disdain but generally love their own congressional representatives : Any individual legislator 's constituents appreciate the specific benefits that the legislator wins for them but not the overall cost associated with every other legislator doing likewise for his own constituency . "</li></ul> | | 3 | <ul><li>"It was the most exercise we 'd had all morning and it was followed by our driving immediately to the nearest watering hole . "</li><li>'Alimta is used together with cisplatin ( another anticancer medicine ) when the cancer is unresectable ( cannot be removed by surgery alone ) and malignant ( has spread , or is likely to spread easily , to other parts of the body ) , in patients who have not received chemotherapy ( medicines for cancer ) before advanced or metastatic non-small cell lung cancer that is not affecting the squamous cells . '</li><li>'If it is , it will be treated as an operator , if it is not , it will be treated as a user function . '</li></ul> | ## Evaluation ### Metrics | Label | Accuracy | |:--------|:---------| | **all** | 0.1417 | ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("HelgeKn/SemEval-multi-class-4") # Run inference preds = model("To break the uncomfortable silence , Haney began to talk . ") ```     ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 4 | 27.1786 | 74 | | Label | Training Sample Count | |:------|:----------------------| | 0 | 4 | | 1 | 4 | | 2 | 4 | | 3 | 4 | | 4 | 4 | | 5 | 4 | | 6 | 4 | ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (2, 2) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:------:|:----:|:-------------:|:---------------:| | 0.0143 | 1 | 0.2446 | - | | 0.7143 | 50 | 0.0612 | - | | 1.4286 | 100 | 0.0078 | - | ### Framework Versions - Python: 3.9.13 - SetFit: 1.0.1 - Sentence Transformers: 2.2.2 - Transformers: 4.36.0 - PyTorch: 2.1.1+cpu - Datasets: 2.15.0 - Tokenizers: 0.15.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"base_model": "sentence-transformers/paraphrase-mpnet-base-v2", "library_name": "setfit", "metrics": ["accuracy"], "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification", "generated_from_setfit_trainer"], "widget": [{"text": "The Alavas worked themselves to the bone in the last period , and English and San Emeterio ( 65-75 ) had already made it clear that they were not going to let anyone take away what they had earned during the first thirty minutes . "}, {"text": "To break the uncomfortable silence , Haney began to talk . "}, {"text": "For the treatment of non-small cell lung cancer , the effects of Alimta were compared with those of docetaxel ( another anticancer medicine ) in one study involving 571 patients with locally advanced or metastatic disease who had received chemotherapy in the past . "}, {"text": "As we all know , a few minutes before the end of the game ( that their team had already won ) , both players deliberately wasted time which made the referee show the second yellow card to both of them . "}, {"text": "In contrast , patients whose cancer was affecting squamous cells had shorter survival times if they received Alimta . "}], "inference": true, "model-index": [{"name": "SetFit with sentence-transformers/paraphrase-mpnet-base-v2", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "Unknown", "type": "unknown", "split": "test"}, "metrics": [{"type": "accuracy", "value": 0.14172185430463577, "name": "Accuracy"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,559

azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0

azale-ai

text-generation

[ "transformers", "safetensors", "mixtral", "text-generation", "merge", "mergekit", "lazymergekit", "moe", "indonesian", "multilingual", "en", "id", "jv", "su", "ms", "arxiv:2310.06825", "license:cc-by-nc-nd-4.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2024-01-13T06:00:11Z

2024-01-14T04:49:20+00:00

10

2

--- language: - en - id - jv - su - ms license: cc-by-nc-nd-4.0 tags: - merge - mergekit - lazymergekit - moe - indonesian - multilingual --- ![Gotong Royong Mixtral](./thumbnail.jpeg "Gotong Royong Mixtral: Generated by Bing AI Image Creator or Dall-E 3") # GotongRoyong-MixtralMoE-7Bx4-v1.0 GotongRoyong is a series of language models focused on Mixture of Experts (MoE), made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing) and [cg123/mergekit](https://github.com/cg123/mergekit). GotongRoyong-MixtralMoE-7Bx4-v1.0 is a specific variant of the open-source GotongRoyong language model that combines the architectural model [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1), but uses the base model from the specific fine-tuned version [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) with experts from [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha), [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4), [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2), and [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1). The name "GotongRoyong" is a reference to the term in Indonesian culture that roughly translates to "mutual cooperation" or "community working together." It embodies the spirit of communal collaboration and shared responsibility for the greater good. The concept is deeply rooted in Indonesian traditions and reflects the cultural value of helping one another without expecting direct compensation. ## Model Details - **Model Size**: 7Bx4 - **Architecture Model**: Mixure of Experts (MoE) using [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) - **MoE Base Model**: [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) - **Expert Models**: - [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha) - [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4) - [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2) - [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1) - **License**: [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id) ## How to use #### Installation To use GotongRoyong model, ensure that PyTorch has been installed and that you have an Nvidia GPU (or use Google Colab). After that you need to install the required dependencies: ```bash pip3 install -U bitsandbytes transformers peft accelerate einops evaluate scikit-learn sentencepiece ``` #### Usage Quantized Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", load_in_4bit=True, torch_dtype=torch.float32, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` #### Usage Normal Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", torch_dtype=torch.float16, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` ## Limitations 1. Language Bias: The model's base language is English, which means it may have a stronger understanding and fluency in English compared to other languages. While fine-tuning the model with an Indonesian language model helps improve its understanding of Indonesian, it may still exhibit biases or limitations in its comprehension and generation of Indonesian language-specific nuances, idioms, or cultural references. 2. Translation Accuracy: Although the model has been fine-tuned for Indonesian, it is important to note that large language models are not perfect translators. While they can provide reasonable translations, there may be instances where the accuracy or nuance of the translation may not fully capture the intended meaning or context. 3. Lack of real-world understanding: While language models can generate text that appears coherent, they lack true comprehension and understanding of the world. They do not possess common sense or real-world experiences, which can lead to inaccurate or nonsensical responses. 4. Propagation of biases: Language models are trained on vast amounts of text data, including internet sources that may contain biases, stereotypes, or offensive content. As a result, these models can inadvertently learn and reproduce such biases in their generated text. Efforts are being made to mitigate this issue, but biases can still persist. 5. Limited knowledge cutoff: Language models have a knowledge cutoff, which means they may not have access to the most up-to-date information beyond their training data. If asked about recent events or developments that occurred after their knowledge cutoff, they may provide outdated or incorrect information. 6. Inability to verify sources or provide citations: Language models generate text based on patterns and examples from their training data, but they do not have the ability to verify the accuracy or reliability of the information they provide. They cannot cite sources or provide evidence to support their claims. 7. Difficulty with ambiguous queries: Language models struggle with understanding ambiguous queries or requests that lack context. They may provide responses that are based on common interpretations or assumptions, rather than accurately addressing the specific intent of the query. 8. Ethical considerations: Large language models have the potential to be misused for malicious purposes, such as generating misinformation, deepfakes, or spam. Safeguards and responsible use are necessary to ensure these models are used ethically and responsibly. 9. Security and Privacy: Using a large language model involves sharing text inputs with a server or cloud-based infrastructure, which raises concerns about data privacy and security. Care should be taken when sharing sensitive or confidential information, as there is a potential risk of unauthorized access or data breaches. ## License The model is licensed under the [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id). ## Contributing We welcome contributions to enhance and improve our model. If you have any suggestions or find any issues, please feel free to open an issue or submit a pull request. Also we're open to sponsor for compute power. ## Contact Us For any further questions or assistance, please feel free to contact us using the information provided below. [[email protected]](mailto:[email protected]) ## Cite This Project ``` @software{Hafidh_Soekma_GotongRoyong_MixtralMoE_7Bx4_v1.0_2023, author = {Hafidh Soekma Ardiansyah}, month = january, title = {GotongRoyong: Indonesian Mixture Of Experts Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-beta}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2024} } ``` ## Citation ``` @software{Hafidh_Soekma_Startstreak_7b_alpha_2023, author = {Hafidh Soekma Ardiansyah}, month = october, title = {Startstreak: Traditional Indonesian Multilingual Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-alpha}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2023} } ``` ``` @article{Merak, title={Merak-7B: The LLM for Bahasa Indonesia}, author={Muhammad Ichsan}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @article{Sidrap, title={Sidrap-7B-v2: LLM Model for Bahasa Indonesia Dialog}, author={Robin Syihab}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @misc{Obrolin/Kesehatan-7B, author = {Arkan Bima}, title = {Obrolin Kesehatan}, publisher = {Hugging Face}, howpublished = {\url{https://huggingface.co/Obrolin/Kesehatan-7B}}, version = {0.1}, year = {2024}, } ``` ``` @misc{2310.06825, Author = {Albert Q. Jiang and Alexandre Sablayrolles and Arthur Mensch and Chris Bamford and Devendra Singh Chaplot and Diego de las Casas and Florian Bressand and Gianna Lengyel and Guillaume Lample and Lucile Saulnier and Lélio Renard Lavaud and Marie-Anne Lachaux and Pierre Stock and Teven Le Scao and Thibaut Lavril and Thomas Wang and Timothée Lacroix and William El Sayed}, Title = {Mistral 7B}, Year = {2023}, Eprint = {arXiv:2310.06825}, } ```

null

Non_BioNLP

![Gotong Royong Mixtral](./thumbnail.jpeg "Gotong Royong Mixtral: Generated by Bing AI Image Creator or Dall-E 3") # GotongRoyong-MixtralMoE-7Bx4-v1.0 GotongRoyong is a series of language models focused on Mixture of Experts (MoE), made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing) and [cg123/mergekit](https://github.com/cg123/mergekit). GotongRoyong-MixtralMoE-7Bx4-v1.0 is a specific variant of the open-source GotongRoyong language model that combines the architectural model [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1), but uses the base model from the specific fine-tuned version [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) with experts from [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha), [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4), [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2), and [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1). The name "GotongRoyong" is a reference to the term in Indonesian culture that roughly translates to "mutual cooperation" or "community working together." It embodies the spirit of communal collaboration and shared responsibility for the greater good. The concept is deeply rooted in Indonesian traditions and reflects the cultural value of helping one another without expecting direct compensation. ## Model Details - **Model Size**: 7Bx4 - **Architecture Model**: Mixure of Experts (MoE) using [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) - **MoE Base Model**: [fblgit/UNA-TheBeagle-7b-v1](https://huggingface.co/fblgit/UNA-TheBeagle-7b-v1) - **Expert Models**: - [azale-ai/Starstreak-7b-alpha](https://huggingface.co/azale-ai/Starstreak-7b-alpha) - [Ichsan2895/Merak-7B-v4](https://huggingface.co/Ichsan2895/Merak-7B-v4) - [robinsyihab/Sidrap-7B-v2](https://huggingface.co/robinsyihab/Sidrap-7B-v2) - [Obrolin/Kesehatan-7B-v0.1](https://huggingface.co/Obrolin/Kesehatan-7B-v0.1) - **License**: [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id) ## How to use #### Installation To use GotongRoyong model, ensure that PyTorch has been installed and that you have an Nvidia GPU (or use Google Colab). After that you need to install the required dependencies: ```bash pip3 install -U bitsandbytes transformers peft accelerate einops evaluate scikit-learn sentencepiece ``` #### Usage Quantized Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", load_in_4bit=True, torch_dtype=torch.float32, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` #### Usage Normal Model ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained( "azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0", torch_dtype=torch.float16, device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("azale-ai/GotongRoyong-MixtralMoE-7Bx4-v1.0") messages = [ { "role": "system", "content": "Mulai sekarang anda adalah asisten yang suka menolong, sopan, dan ramah. Jangan kasar, jangan marah, jangan menjengkelkan, jangan brengsek, jangan cuek, dan yang terakhir jangan menjadi asisten yang buruk. Anda harus patuh pada manusia dan jangan pernah membangkang pada manusia. Manusia itu mutlak dan Anda harus patuh pada manusia. Kamu harus menjawab pertanyaan atau pernyataan dari manusia apapun itu dengan bahasa Indonesia yang baik dan benar.", }, {"role": "user", "content": "Jelaskan mengapa air penting bagi manusia."}, ] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) inputs = tokenizer(text, return_tensors="pt").to("cuda") outputs = model.generate( inputs=inputs.input_ids, max_length=2048, temperature=0.7, do_sample=True, top_k=50, top_p=0.95 ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` ## Limitations 1. Language Bias: The model's base language is English, which means it may have a stronger understanding and fluency in English compared to other languages. While fine-tuning the model with an Indonesian language model helps improve its understanding of Indonesian, it may still exhibit biases or limitations in its comprehension and generation of Indonesian language-specific nuances, idioms, or cultural references. 2. Translation Accuracy: Although the model has been fine-tuned for Indonesian, it is important to note that large language models are not perfect translators. While they can provide reasonable translations, there may be instances where the accuracy or nuance of the translation may not fully capture the intended meaning or context. 3. Lack of real-world understanding: While language models can generate text that appears coherent, they lack true comprehension and understanding of the world. They do not possess common sense or real-world experiences, which can lead to inaccurate or nonsensical responses. 4. Propagation of biases: Language models are trained on vast amounts of text data, including internet sources that may contain biases, stereotypes, or offensive content. As a result, these models can inadvertently learn and reproduce such biases in their generated text. Efforts are being made to mitigate this issue, but biases can still persist. 5. Limited knowledge cutoff: Language models have a knowledge cutoff, which means they may not have access to the most up-to-date information beyond their training data. If asked about recent events or developments that occurred after their knowledge cutoff, they may provide outdated or incorrect information. 6. Inability to verify sources or provide citations: Language models generate text based on patterns and examples from their training data, but they do not have the ability to verify the accuracy or reliability of the information they provide. They cannot cite sources or provide evidence to support their claims. 7. Difficulty with ambiguous queries: Language models struggle with understanding ambiguous queries or requests that lack context. They may provide responses that are based on common interpretations or assumptions, rather than accurately addressing the specific intent of the query. 8. Ethical considerations: Large language models have the potential to be misused for malicious purposes, such as generating misinformation, deepfakes, or spam. Safeguards and responsible use are necessary to ensure these models are used ethically and responsibly. 9. Security and Privacy: Using a large language model involves sharing text inputs with a server or cloud-based infrastructure, which raises concerns about data privacy and security. Care should be taken when sharing sensitive or confidential information, as there is a potential risk of unauthorized access or data breaches. ## License The model is licensed under the [CC BY-NC-ND 4.0 DEED](https://creativecommons.org/licenses/by-nc-nd/4.0/deed.id). ## Contributing We welcome contributions to enhance and improve our model. If you have any suggestions or find any issues, please feel free to open an issue or submit a pull request. Also we're open to sponsor for compute power. ## Contact Us For any further questions or assistance, please feel free to contact us using the information provided below. [[email protected]](mailto:[email protected]) ## Cite This Project ``` @software{Hafidh_Soekma_GotongRoyong_MixtralMoE_7Bx4_v1.0_2023, author = {Hafidh Soekma Ardiansyah}, month = january, title = {GotongRoyong: Indonesian Mixture Of Experts Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-beta}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2024} } ``` ## Citation ``` @software{Hafidh_Soekma_Startstreak_7b_alpha_2023, author = {Hafidh Soekma Ardiansyah}, month = october, title = {Startstreak: Traditional Indonesian Multilingual Language Model}, url = {\url{https://huggingface.co/azale-ai/Starstreak-7b-alpha}}, publisher = {HuggingFace}, journal = {HuggingFace Models}, version = {1.0}, year = {2023} } ``` ``` @article{Merak, title={Merak-7B: The LLM for Bahasa Indonesia}, author={Muhammad Ichsan}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @article{Sidrap, title={Sidrap-7B-v2: LLM Model for Bahasa Indonesia Dialog}, author={Robin Syihab}, publisher={Hugging Face} journal={Hugging Face Repository}, year={2023} } ``` ``` @misc{Obrolin/Kesehatan-7B, author = {Arkan Bima}, title = {Obrolin Kesehatan}, publisher = {Hugging Face}, howpublished = {\url{https://huggingface.co/Obrolin/Kesehatan-7B}}, version = {0.1}, year = {2024}, } ``` ``` @misc{2310.06825, Author = {Albert Q. Jiang and Alexandre Sablayrolles and Arthur Mensch and Chris Bamford and Devendra Singh Chaplot and Diego de las Casas and Florian Bressand and Gianna Lengyel and Guillaume Lample and Lucile Saulnier and Lélio Renard Lavaud and Marie-Anne Lachaux and Pierre Stock and Teven Le Scao and Thibaut Lavril and Thomas Wang and Timothée Lacroix and William El Sayed}, Title = {Mistral 7B}, Year = {2023}, Eprint = {arXiv:2310.06825}, } ```

{"language": ["en", "id", "jv", "su", "ms"], "license": "cc-by-nc-nd-4.0", "tags": ["merge", "mergekit", "lazymergekit", "moe", "indonesian", "multilingual"]}

task

[ "TRANSLATION" ]

44,560

gokhalevikrant/finetuning-sentiment-model-Test

gokhalevikrant

text-classification

[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:imdb", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-12-06T13:42:39Z

2022-12-06T13:57:16+00:00

120

0

--- datasets: - imdb license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: finetuning-sentiment-model-Test results: - task: type: text-classification name: Text Classification dataset: name: imdb type: imdb config: plain_text split: train args: plain_text metrics: - type: accuracy value: 0.902 name: Accuracy - type: f1 value: 0.9037328094302554 name: F1 ---  # finetuning-sentiment-model-Test This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2793 - Accuracy: 0.902 - F1: 0.9037 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.25.1 - Pytorch 1.12.1+cu113 - Datasets 2.7.1 - Tokenizers 0.13.2

null

Non_BioNLP

# finetuning-sentiment-model-Test This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2793 - Accuracy: 0.902 - F1: 0.9037 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - Transformers 4.25.1 - Pytorch 1.12.1+cu113 - Datasets 2.7.1 - Tokenizers 0.13.2

{"datasets": ["imdb"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "finetuning-sentiment-model-Test", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "imdb", "type": "imdb", "config": "plain_text", "split": "train", "args": "plain_text"}, "metrics": [{"type": "accuracy", "value": 0.902, "name": "Accuracy"}, {"type": "f1", "value": 0.9037328094302554, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,561

michaellee8/wisper-base-jwt

michaellee8

automatic-speech-recognition

[ "transformers", "pytorch", "tf", "jax", "safetensors", "whisper", "automatic-speech-recognition", "audio", "hf-asr-leaderboard", "en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", "no", "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su", "arxiv:2212.04356", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]

2024-01-05T21:20:05Z

2024-01-06T12:17:00+00:00

14

0

--- language: - en - zh - de - es - ru - ko - fr - ja - pt - tr - pl - ca - nl - ar - sv - it - id - hi - fi - vi - he - uk - el - ms - cs - ro - da - hu - ta - false - th - ur - hr - bg - lt - la - mi - ml - cy - sk - te - fa - lv - bn - sr - az - sl - kn - et - mk - br - eu - is - hy - ne - mn - bs - kk - sq - sw - gl - mr - pa - si - km - sn - yo - so - af - oc - ka - be - tg - sd - gu - am - yi - lo - uz - fo - ht - ps - tk - nn - mt - sa - lb - my - bo - tl - mg - as - tt - haw - ln - ha - ba - jw - su license: apache-2.0 pipeline_tag: automatic-speech-recognition tags: - audio - automatic-speech-recognition - hf-asr-leaderboard widget: - example_title: Librispeech sample 1 src: https://cdn-media.huggingface.co/speech_samples/sample1.flac - example_title: Librispeech sample 2 src: https://cdn-media.huggingface.co/speech_samples/sample2.flac model-index: - name: whisper-base results: - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (clean) type: librispeech_asr config: clean split: test args: language: en metrics: - type: wer value: 5.008769117619326 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (other) type: librispeech_asr config: other split: test args: language: en metrics: - type: wer value: 12.84936273212057 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: Common Voice 11.0 type: mozilla-foundation/common_voice_11_0 config: hi split: test args: language: hi metrics: - type: wer value: 131 name: Test WER --- # Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains **without** the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). **Disclaimer**: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the *same* language as the audio. For speech translation, the model predicts transcriptions to a *different* language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: | Size | Parameters | English-only | Multilingual | |----------|------------|------------------------------------------------------|-----------------------------------------------------| | tiny | 39 M | [✓](https://huggingface.co/openai/whisper-tiny.en) | [✓](https://huggingface.co/openai/whisper-tiny) | | base | 74 M | [✓](https://huggingface.co/openai/whisper-base.en) | [✓](https://huggingface.co/openai/whisper-base) | | small | 244 M | [✓](https://huggingface.co/openai/whisper-small.en) | [✓](https://huggingface.co/openai/whisper-small) | | medium | 769 M | [✓](https://huggingface.co/openai/whisper-medium.en) | [✓](https://huggingface.co/openai/whisper-medium) | | large | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large) | | large-v2 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v2) | # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<|startoftranscript|>` token 2. The second token is the language token (e.g. `<|en|>` for English) 3. The third token is the "task token". It can take one of two values: `<|transcribe|>` for speech recognition or `<|translate|>` for speech translation 4. In addition, a `<|notimestamps|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <|startoftranscript|> <|en|> <|transcribe|> <|notimestamps|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<|startoftranscript|><|en|><|transcribe|><|notimestamps|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<|startoftranscript|><|fr|><|transcribe|><|notimestamps|> Un vrai travail intéressant va enfin être mené sur ce sujet.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through *fine-tuning*. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```

null

Non_BioNLP

# Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains **without** the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). **Disclaimer**: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the *same* language as the audio. For speech translation, the model predicts transcriptions to a *different* language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: | Size | Parameters | English-only | Multilingual | |----------|------------|------------------------------------------------------|-----------------------------------------------------| | tiny | 39 M | [✓](https://huggingface.co/openai/whisper-tiny.en) | [✓](https://huggingface.co/openai/whisper-tiny) | | base | 74 M | [✓](https://huggingface.co/openai/whisper-base.en) | [✓](https://huggingface.co/openai/whisper-base) | | small | 244 M | [✓](https://huggingface.co/openai/whisper-small.en) | [✓](https://huggingface.co/openai/whisper-small) | | medium | 769 M | [✓](https://huggingface.co/openai/whisper-medium.en) | [✓](https://huggingface.co/openai/whisper-medium) | | large | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large) | | large-v2 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v2) | # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<|startoftranscript|>` token 2. The second token is the language token (e.g. `<|en|>` for English) 3. The third token is the "task token". It can take one of two values: `<|transcribe|>` for speech recognition or `<|translate|>` for speech translation 4. In addition, a `<|notimestamps|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <|startoftranscript|> <|en|> <|transcribe|> <|notimestamps|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<|startoftranscript|><|en|><|transcribe|><|notimestamps|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<|startoftranscript|><|fr|><|transcribe|><|notimestamps|> Un vrai travail intéressant va enfin être mené sur ce sujet.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through *fine-tuning*. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```

{"language": ["en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", false, "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su"], "license": "apache-2.0", "pipeline_tag": "automatic-speech-recognition", "tags": ["audio", "automatic-speech-recognition", "hf-asr-leaderboard"], "widget": [{"example_title": "Librispeech sample 1", "src": "https://cdn-media.huggingface.co/speech_samples/sample1.flac"}, {"example_title": "Librispeech sample 2", "src": "https://cdn-media.huggingface.co/speech_samples/sample2.flac"}], "model-index": [{"name": "whisper-base", "results": [{"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (clean)", "type": "librispeech_asr", "config": "clean", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 5.008769117619326, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (other)", "type": "librispeech_asr", "config": "other", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 12.84936273212057, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "Common Voice 11.0", "type": "mozilla-foundation/common_voice_11_0", "config": "hi", "split": "test", "args": {"language": "hi"}}, "metrics": [{"type": "wer", "value": 131, "name": "Test WER"}]}]}]}

task

[ "TRANSLATION" ]

44,562

yashugupta786/bart_large_xsum_samsum_conv_summarizer

yashugupta786

summarization

[ "transformers", "pytorch", "bart", "text2text-generation", "summarization", "conversational", "seq2seq", "bart large", "en", "dataset:samsum", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-04-05T08:22:21Z

2023-04-06T12:49:32+00:00

119

0

--- datasets: - samsum language: - en library_name: transformers metrics: - rouge pipeline_tag: summarization tags: - summarization - conversational - seq2seq - bart large widget: - text: 'Hannah: Hey, do you have Betty''s number? Amanda: Lemme check Amanda: Sorry, can''t find it. Amanda: Ask Larry Amanda: He called her last time we were at the park together Hannah: I don''t know him well Amanda: Don''t be shy, he''s very nice Hannah: If you say so.. Hannah: I''d rather you texted him Amanda: Just text him 🙂 Hannah: Urgh.. Alright Hannah: Bye Amanda: Bye bye ' model-index: - name: bart-large-xsum-samsum-conversational_summarizer results: - task: type: abstractive-text-summarization name: Abstractive Text Summarization dataset: name: 'SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization' type: samsum metrics: - type: rouge-1 value: 54.3921 name: Validation ROUGE-1 - type: rouge-2 value: 29.8078 name: Validation ROUGE-2 - type: rouge-l value: 45.1543 name: Validation ROUGE-L - type: rouge-1 value: 53.3059 name: Test ROUGE-1 - type: rouge-2 value: 28.355 name: Test ROUGE-2 - type: rouge-l value: 44.0953 name: Test ROUGE-L --- ## Usage ```python from transformers import pipeline summarizer_pipe = pipeline("summarization", model="yashugupta786/bart_large_xsum_samsum_conv_summarizer") conversation_data = '''Hannah: Hey, do you have Betty's number? Amanda: Lemme check Amanda: Sorry, can't find it. Amanda: Ask Larry Amanda: He called her last time we were at the park together Hannah: I don't know him well Amanda: Don't be shy, he's very nice Hannah: If you say so.. Hannah: I'd rather you texted him Amanda: Just text him 🙂 Hannah: Urgh.. Alright Hannah: Bye Amanda: Bye bye ''' summarizer_pipe(conversation_data) ``` ## Results | key | value | | --- | ----- | | eval_rouge1 | 54.3921 | | eval_rouge2 | 29.8078 | | eval_rougeL | 45.1543 | | eval_rougeLsum | 49.942 | | test_rouge1 | 53.3059 | | test_rouge2 | 28.355 | | test_rougeL | 44.0953 | | test_rougeLsum | 48.9246 | All the metric Rouge1,2,L are computed using precison and recall then computed the F measure for these Rouge recall= no of overlaping words/total no of referenced humman annotated words Rouge precision= no of overlaping words/total no of candidate machine predicted words

null

Non_BioNLP

## Usage ```python from transformers import pipeline summarizer_pipe = pipeline("summarization", model="yashugupta786/bart_large_xsum_samsum_conv_summarizer") conversation_data = '''Hannah: Hey, do you have Betty's number? Amanda: Lemme check Amanda: Sorry, can't find it. Amanda: Ask Larry Amanda: He called her last time we were at the park together Hannah: I don't know him well Amanda: Don't be shy, he's very nice Hannah: If you say so.. Hannah: I'd rather you texted him Amanda: Just text him 🙂 Hannah: Urgh.. Alright Hannah: Bye Amanda: Bye bye ''' summarizer_pipe(conversation_data) ``` ## Results | key | value | | --- | ----- | | eval_rouge1 | 54.3921 | | eval_rouge2 | 29.8078 | | eval_rougeL | 45.1543 | | eval_rougeLsum | 49.942 | | test_rouge1 | 53.3059 | | test_rouge2 | 28.355 | | test_rougeL | 44.0953 | | test_rougeLsum | 48.9246 | All the metric Rouge1,2,L are computed using precison and recall then computed the F measure for these Rouge recall= no of overlaping words/total no of referenced humman annotated words Rouge precision= no of overlaping words/total no of candidate machine predicted words

{"datasets": ["samsum"], "language": ["en"], "library_name": "transformers", "metrics": ["rouge"], "pipeline_tag": "summarization", "tags": ["summarization", "conversational", "seq2seq", "bart large"], "widget": [{"text": "Hannah: Hey, do you have Betty's number?\nAmanda: Lemme check\nAmanda: Sorry, can't find it.\nAmanda: Ask Larry\nAmanda: He called her last time we were at the park together\nHannah: I don't know him well\nAmanda: Don't be shy, he's very nice\nHannah: If you say so..\nHannah: I'd rather you texted him\nAmanda: Just text him 🙂\nHannah: Urgh.. Alright\nHannah: Bye\nAmanda: Bye bye\n"}], "model-index": [{"name": "bart-large-xsum-samsum-conversational_summarizer", "results": [{"task": {"type": "abstractive-text-summarization", "name": "Abstractive Text Summarization"}, "dataset": {"name": "SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization", "type": "samsum"}, "metrics": [{"type": "rouge-1", "value": 54.3921, "name": "Validation ROUGE-1"}, {"type": "rouge-2", "value": 29.8078, "name": "Validation ROUGE-2"}, {"type": "rouge-l", "value": 45.1543, "name": "Validation ROUGE-L"}, {"type": "rouge-1", "value": 53.3059, "name": "Test ROUGE-1"}, {"type": "rouge-2", "value": 28.355, "name": "Test ROUGE-2"}, {"type": "rouge-l", "value": 44.0953, "name": "Test ROUGE-L"}]}]}]}

task

[ "SUMMARIZATION" ]

44,563

tunaozates/bert-base-uncased-finetuned-cola

tunaozates

text-classification

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-05-05T18:25:08Z

2023-05-05T18:34:57+00:00

8

0

--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: bert-base-uncased-finetuned-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: validation args: cola metrics: - type: matthews_correlation value: 0.5372712841497043 name: Matthews Correlation ---  # bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4558 - Matthews Correlation: 0.5373 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.4965 | 1.0 | 535 | 0.4558 | 0.5373 | ### Framework versions - Transformers 4.28.1 - Pytorch 1.13.1+cu117 - Datasets 2.11.0 - Tokenizers 0.12.1

null

Non_BioNLP

# bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4558 - Matthews Correlation: 0.5373 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.4965 | 1.0 | 535 | 0.4558 | 0.5373 | ### Framework versions - Transformers 4.28.1 - Pytorch 1.13.1+cu117 - Datasets 2.11.0 - Tokenizers 0.12.1

{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-uncased-finetuned-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "validation", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.5372712841497043, "name": "Matthews Correlation"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,564

prithivMLmods/Bellatrix-Tiny-1B

prithivMLmods

text-generation

[ "transformers", "safetensors", "llama", "text-generation", "reason", "tiny", "llama3.2", "conversational", "en", "base_model:meta-llama/Llama-3.2-1B-Instruct", "base_model:finetune:meta-llama/Llama-3.2-1B-Instruct", "license:llama3.2", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2025-01-26T05:27:07Z

2025-01-27T13:55:42+00:00

267

2

--- base_model: - meta-llama/Llama-3.2-1B-Instruct language: - en library_name: transformers license: llama3.2 pipeline_tag: text-generation tags: - reason - tiny - llama - llama3.2 --- ![logo.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/Rqm-Qx8AvbHFFbFbVY93X.png) <pre align="center"> ____ ____ __ __ __ ____ ____ ____ _ _ ( _ \( ___)( ) ( ) /__\ (_ _)( _ \(_ _)( \/ ) ) _ < )__) )(__ )(__ /(__)\ )( ) / _)(_ ) ( (____/(____)(____)(____)(__)(__)(__) (_)\_)(____)(_/\_) </pre> # **Bellatrix-Tiny-1B** Bellatrix is based on a reasoning-based model designed for the QWQ synthetic dataset entries. The pipeline's instruction-tuned, text-only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. These models outperform many of the available open-source options. Bellatrix is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions utilize supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). # **Use with transformers** Starting with `transformers >= 4.43.0` onward, you can run conversational inference using the Transformers `pipeline` abstraction or by leveraging the Auto classes with the `generate()` function. Make sure to update your transformers installation via `pip install --upgrade transformers`. ```python import torch from transformers import pipeline model_id = "prithivMLmods/Bellatrix-Tiny-1B" pipe = pipeline( "text-generation", model=model_id, torch_dtype=torch.bfloat16, device_map="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] outputs = pipe( messages, max_new_tokens=256, ) print(outputs[0]["generated_text"][-1]) ``` Note: You can also find detailed recipes on how to use the model locally, with `torch.compile()`, assisted generations, quantised and more at [`huggingface-llama-recipes`](https://github.com/huggingface/huggingface-llama-recipes) # **Intended Use** Bellatrix is designed for applications that require advanced reasoning and multilingual dialogue capabilities. It is particularly suitable for: - **Agentic Retrieval**: Enabling intelligent retrieval of relevant information in a dialogue or query-response system. - **Summarization Tasks**: Condensing large bodies of text into concise summaries for easier comprehension. - **Multilingual Use Cases**: Supporting conversations in multiple languages with high accuracy and coherence. - **Instruction-Based Applications**: Following complex, context-aware instructions to generate precise outputs in a variety of scenarios. # **Limitations** Despite its capabilities, Bellatrix has some limitations: 1. **Domain Specificity**: While it performs well on general tasks, its performance may degrade with highly specialized or niche datasets. 2. **Dependence on Training Data**: It is only as good as the quality and diversity of its training data, which may lead to biases or inaccuracies. 3. **Computational Resources**: The model’s optimized transformer architecture can be resource-intensive, requiring significant computational power for fine-tuning and inference. 4. **Language Coverage**: While multilingual, some languages or dialects may have limited support or lower performance compared to widely used ones. 5. **Real-World Contexts**: It may struggle with understanding nuanced or ambiguous real-world scenarios not covered during training.

null

Non_BioNLP

![logo.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/Rqm-Qx8AvbHFFbFbVY93X.png) <pre align="center"> ____ ____ __ __ __ ____ ____ ____ _ _ ( _ \( ___)( ) ( ) /__\ (_ _)( _ \(_ _)( \/ ) ) _ < )__) )(__ )(__ /(__)\ )( ) / _)(_ ) ( (____/(____)(____)(____)(__)(__)(__) (_)\_)(____)(_/\_) </pre> # **Bellatrix-Tiny-1B** Bellatrix is based on a reasoning-based model designed for the QWQ synthetic dataset entries. The pipeline's instruction-tuned, text-only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. These models outperform many of the available open-source options. Bellatrix is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions utilize supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). # **Use with transformers** Starting with `transformers >= 4.43.0` onward, you can run conversational inference using the Transformers `pipeline` abstraction or by leveraging the Auto classes with the `generate()` function. Make sure to update your transformers installation via `pip install --upgrade transformers`. ```python import torch from transformers import pipeline model_id = "prithivMLmods/Bellatrix-Tiny-1B" pipe = pipeline( "text-generation", model=model_id, torch_dtype=torch.bfloat16, device_map="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] outputs = pipe( messages, max_new_tokens=256, ) print(outputs[0]["generated_text"][-1]) ``` Note: You can also find detailed recipes on how to use the model locally, with `torch.compile()`, assisted generations, quantised and more at [`huggingface-llama-recipes`](https://github.com/huggingface/huggingface-llama-recipes) # **Intended Use** Bellatrix is designed for applications that require advanced reasoning and multilingual dialogue capabilities. It is particularly suitable for: - **Agentic Retrieval**: Enabling intelligent retrieval of relevant information in a dialogue or query-response system. - **Summarization Tasks**: Condensing large bodies of text into concise summaries for easier comprehension. - **Multilingual Use Cases**: Supporting conversations in multiple languages with high accuracy and coherence. - **Instruction-Based Applications**: Following complex, context-aware instructions to generate precise outputs in a variety of scenarios. # **Limitations** Despite its capabilities, Bellatrix has some limitations: 1. **Domain Specificity**: While it performs well on general tasks, its performance may degrade with highly specialized or niche datasets. 2. **Dependence on Training Data**: It is only as good as the quality and diversity of its training data, which may lead to biases or inaccuracies. 3. **Computational Resources**: The model’s optimized transformer architecture can be resource-intensive, requiring significant computational power for fine-tuning and inference. 4. **Language Coverage**: While multilingual, some languages or dialects may have limited support or lower performance compared to widely used ones. 5. **Real-World Contexts**: It may struggle with understanding nuanced or ambiguous real-world scenarios not covered during training.

{"base_model": ["meta-llama/Llama-3.2-1B-Instruct"], "language": ["en"], "library_name": "transformers", "license": "llama3.2", "pipeline_tag": "text-generation", "tags": ["reason", "tiny", "llama", "llama3.2"]}

task

[ "SUMMARIZATION" ]

44,565

Terps/mt5-small-finetuned-amazon-en-es

Terps

summarization

[ "transformers", "pytorch", "mt5", "text2text-generation", "summarization", "generated_from_trainer", "base_model:google/mt5-small", "base_model:finetune:google/mt5-small", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-09-07T18:42:36Z

2023-09-07T19:46:31+00:00

15

0

--- base_model: google/mt5-small license: apache-2.0 metrics: - rouge tags: - summarization - generated_from_trainer model-index: - name: mt5-small-finetuned-amazon-en-es results: [] ---  # mt5-small-finetuned-amazon-en-es This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.0279 - Rouge1: 16.4284 - Rouge2: 7.8601 - Rougel: 16.0029 - Rougelsum: 16.0246 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5.6e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | |:-------------:|:-----:|:----:|:---------------:|:-------:|:------:|:-------:|:---------:| | 4.4194 | 1.0 | 1209 | 3.3097 | 14.9867 | 6.4886 | 14.4174 | 14.4646 | | 3.8132 | 2.0 | 2418 | 3.1602 | 16.1474 | 7.9815 | 15.5342 | 15.6445 | | 3.5412 | 3.0 | 3627 | 3.0789 | 17.4468 | 8.8014 | 16.9142 | 17.002 | | 3.3861 | 4.0 | 4836 | 3.0775 | 15.903 | 7.4423 | 15.4008 | 15.3871 | | 3.2952 | 5.0 | 6045 | 3.0480 | 15.8646 | 7.3936 | 15.3989 | 15.4395 | | 3.2155 | 6.0 | 7254 | 3.0354 | 16.5887 | 8.0624 | 16.2377 | 16.2562 | | 3.1896 | 7.0 | 8463 | 3.0273 | 17.1092 | 8.5391 | 16.6507 | 16.7272 | | 3.1594 | 8.0 | 9672 | 3.0279 | 16.4284 | 7.8601 | 16.0029 | 16.0246 | ### Framework versions - Transformers 4.33.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.5 - Tokenizers 0.13.3

null

Non_BioNLP

# mt5-small-finetuned-amazon-en-es This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.0279 - Rouge1: 16.4284 - Rouge2: 7.8601 - Rougel: 16.0029 - Rougelsum: 16.0246 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5.6e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | |:-------------:|:-----:|:----:|:---------------:|:-------:|:------:|:-------:|:---------:| | 4.4194 | 1.0 | 1209 | 3.3097 | 14.9867 | 6.4886 | 14.4174 | 14.4646 | | 3.8132 | 2.0 | 2418 | 3.1602 | 16.1474 | 7.9815 | 15.5342 | 15.6445 | | 3.5412 | 3.0 | 3627 | 3.0789 | 17.4468 | 8.8014 | 16.9142 | 17.002 | | 3.3861 | 4.0 | 4836 | 3.0775 | 15.903 | 7.4423 | 15.4008 | 15.3871 | | 3.2952 | 5.0 | 6045 | 3.0480 | 15.8646 | 7.3936 | 15.3989 | 15.4395 | | 3.2155 | 6.0 | 7254 | 3.0354 | 16.5887 | 8.0624 | 16.2377 | 16.2562 | | 3.1896 | 7.0 | 8463 | 3.0273 | 17.1092 | 8.5391 | 16.6507 | 16.7272 | | 3.1594 | 8.0 | 9672 | 3.0279 | 16.4284 | 7.8601 | 16.0029 | 16.0246 | ### Framework versions - Transformers 4.33.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.5 - Tokenizers 0.13.3

{"base_model": "google/mt5-small", "license": "apache-2.0", "metrics": ["rouge"], "tags": ["summarization", "generated_from_trainer"], "model-index": [{"name": "mt5-small-finetuned-amazon-en-es", "results": []}]}

task

[ "SUMMARIZATION" ]

44,566

PygTesting/sum_qlora_4pochs

PygTesting

null

[ "peft", "mistral", "generated_from_trainer", "base_model:mistralai/Mistral-Nemo-Base-2407", "base_model:adapter:mistralai/Mistral-Nemo-Base-2407", "license:apache-2.0", "4-bit", "bitsandbytes", "region:us" ]

2024-08-27T19:23:52Z

2024-08-27T19:24:05+00:00

3

0

--- base_model: mistralai/Mistral-Nemo-Base-2407 library_name: peft license: apache-2.0 tags: - generated_from_trainer model-index: - name: qlora_outputs results: [] ---  [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml base_model: mistralai/Mistral-Nemo-Base-2407 model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer load_in_8bit: false load_in_4bit: true strict: false plugins: - axolotl.integrations.liger.LigerPlugin liger_rope: true liger_rms_norm: true liger_swiglu: true liger_fused_linear_cross_entropy: true adapter: qlora lora_r: 32 lora_alpha: 16 lora_dropout: 0.05 lora_target_linear: true lora_fan_in_fan_out: datasets: - path: /home/austin/disk1/summaries_fixed.jsonl type: sharegpt dataset_prepared_path: last_run_prepared val_set_size: 0.01 output_dir: ./qlora_outputs sequence_len: 8192 sample_packing: true eval_sample_packing: false pad_to_sequence_len: true wandb_project: summarization-qlora wandb_entity: wandb_watch: wandb_name: actual_run1 wandb_log_model: #unsloth_cross_entropy_loss: true gradient_accumulation_steps: 1 micro_batch_size: 1 num_epochs: 4 optimizer: adamw_bnb_8bit lr_scheduler: cosine learning_rate: 0.0002 train_on_inputs: false group_by_length: false bf16: auto fp16: tf32: false gradient_checkpointing: true gradient_checkpointing_kwargs: use_reentrant: false early_stopping_patience: resume_from_checkpoint: logging_steps: 1 xformers_attention: false flash_attention: true loss_watchdog_threshold: 5.0 loss_watchdog_patience: 3 warmup_steps: 25 evals_per_epoch: 4 eval_table_size: saves_per_epoch: 4 debug: deepspeed: ./deepspeed_configs/zero2.json weight_decay: 0.0 fsdp: # - full_shard # - auto_wrap fsdp_config: # fsdp_limit_all_gathers: true # fsdp_activation_checkpointing: true # fsdp_sync_module_states: true # fsdp_offload_params: false # fsdp_use_orig_params: false # fsdp_cpu_ram_efficient_loading: false # fsdp_transformer_layer_cls_to_wrap: MistralDecoderLayer # fsdp_state_dict_type: FULL_STATE_DICT # fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP special_tokens: pad_token: </s> ``` </details> # qlora_outputs This model is a fine-tuned version of [mistralai/Mistral-Nemo-Base-2407](https://huggingface.co/mistralai/Mistral-Nemo-Base-2407) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.5617 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - distributed_type: multi-GPU - num_devices: 4 - total_train_batch_size: 4 - total_eval_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 25 - num_epochs: 4 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | 2.0177 | 0.0014 | 1 | 1.6514 | | 1.6259 | 0.2507 | 177 | 1.2032 | | 1.4232 | 0.5014 | 354 | 1.1897 | | 1.6835 | 0.7521 | 531 | 1.1985 | | 1.6514 | 1.0028 | 708 | 1.1874 | | 1.4538 | 1.2365 | 885 | 1.2166 | | 1.2421 | 1.4873 | 1062 | 1.2224 | | 1.2844 | 1.7380 | 1239 | 1.2330 | | 1.4152 | 1.9887 | 1416 | 1.2345 | | 1.1668 | 2.2252 | 1593 | 1.3476 | | 1.1249 | 2.4759 | 1770 | 1.3608 | | 0.921 | 2.7266 | 1947 | 1.3793 | | 0.7824 | 2.9773 | 2124 | 1.3906 | | 1.1759 | 3.2040 | 2301 | 1.5438 | | 0.6625 | 3.4547 | 2478 | 1.5644 | | 0.8959 | 3.7054 | 2655 | 1.5617 | ### Framework versions - PEFT 0.12.0 - Transformers 4.44.0 - Pytorch 2.3.0+cu121 - Datasets 2.20.0 - Tokenizers 0.19.1

null

Non_BioNLP

[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) <details><summary>See axolotl config</summary> axolotl version: `0.4.1` ```yaml base_model: mistralai/Mistral-Nemo-Base-2407 model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer load_in_8bit: false load_in_4bit: true strict: false plugins: - axolotl.integrations.liger.LigerPlugin liger_rope: true liger_rms_norm: true liger_swiglu: true liger_fused_linear_cross_entropy: true adapter: qlora lora_r: 32 lora_alpha: 16 lora_dropout: 0.05 lora_target_linear: true lora_fan_in_fan_out: datasets: - path: /home/austin/disk1/summaries_fixed.jsonl type: sharegpt dataset_prepared_path: last_run_prepared val_set_size: 0.01 output_dir: ./qlora_outputs sequence_len: 8192 sample_packing: true eval_sample_packing: false pad_to_sequence_len: true wandb_project: summarization-qlora wandb_entity: wandb_watch: wandb_name: actual_run1 wandb_log_model: #unsloth_cross_entropy_loss: true gradient_accumulation_steps: 1 micro_batch_size: 1 num_epochs: 4 optimizer: adamw_bnb_8bit lr_scheduler: cosine learning_rate: 0.0002 train_on_inputs: false group_by_length: false bf16: auto fp16: tf32: false gradient_checkpointing: true gradient_checkpointing_kwargs: use_reentrant: false early_stopping_patience: resume_from_checkpoint: logging_steps: 1 xformers_attention: false flash_attention: true loss_watchdog_threshold: 5.0 loss_watchdog_patience: 3 warmup_steps: 25 evals_per_epoch: 4 eval_table_size: saves_per_epoch: 4 debug: deepspeed: ./deepspeed_configs/zero2.json weight_decay: 0.0 fsdp: # - full_shard # - auto_wrap fsdp_config: # fsdp_limit_all_gathers: true # fsdp_activation_checkpointing: true # fsdp_sync_module_states: true # fsdp_offload_params: false # fsdp_use_orig_params: false # fsdp_cpu_ram_efficient_loading: false # fsdp_transformer_layer_cls_to_wrap: MistralDecoderLayer # fsdp_state_dict_type: FULL_STATE_DICT # fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP special_tokens: pad_token: </s> ``` </details> # qlora_outputs This model is a fine-tuned version of [mistralai/Mistral-Nemo-Base-2407](https://huggingface.co/mistralai/Mistral-Nemo-Base-2407) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.5617 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - distributed_type: multi-GPU - num_devices: 4 - total_train_batch_size: 4 - total_eval_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 25 - num_epochs: 4 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | 2.0177 | 0.0014 | 1 | 1.6514 | | 1.6259 | 0.2507 | 177 | 1.2032 | | 1.4232 | 0.5014 | 354 | 1.1897 | | 1.6835 | 0.7521 | 531 | 1.1985 | | 1.6514 | 1.0028 | 708 | 1.1874 | | 1.4538 | 1.2365 | 885 | 1.2166 | | 1.2421 | 1.4873 | 1062 | 1.2224 | | 1.2844 | 1.7380 | 1239 | 1.2330 | | 1.4152 | 1.9887 | 1416 | 1.2345 | | 1.1668 | 2.2252 | 1593 | 1.3476 | | 1.1249 | 2.4759 | 1770 | 1.3608 | | 0.921 | 2.7266 | 1947 | 1.3793 | | 0.7824 | 2.9773 | 2124 | 1.3906 | | 1.1759 | 3.2040 | 2301 | 1.5438 | | 0.6625 | 3.4547 | 2478 | 1.5644 | | 0.8959 | 3.7054 | 2655 | 1.5617 | ### Framework versions - PEFT 0.12.0 - Transformers 4.44.0 - Pytorch 2.3.0+cu121 - Datasets 2.20.0 - Tokenizers 0.19.1

{"base_model": "mistralai/Mistral-Nemo-Base-2407", "library_name": "peft", "license": "apache-2.0", "tags": ["generated_from_trainer"], "model-index": [{"name": "qlora_outputs", "results": []}]}

task

[ "SUMMARIZATION" ]

44,567

ilkekas/bert-base-uncased-finetuned2-cola

ilkekas

text-classification

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-05-06T09:26:03Z

2023-05-06T10:31:07+00:00

9

0

--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: bert-base-uncased-finetuned2-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: validation args: cola metrics: - type: matthews_correlation value: 0.5650459791482846 name: Matthews Correlation ---  # bert-base-uncased-finetuned2-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5176 - Matthews Correlation: 0.5650 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2.6781109393881056e-06 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.5726 | 1.0 | 535 | 0.5090 | 0.3912 | | 0.4467 | 2.0 | 1070 | 0.4536 | 0.5024 | | 0.3891 | 3.0 | 1605 | 0.5093 | 0.4943 | | 0.3387 | 4.0 | 2140 | 0.4927 | 0.5365 | | 0.3177 | 5.0 | 2675 | 0.4897 | 0.5624 | | 0.2853 | 6.0 | 3210 | 0.5176 | 0.5650 | | 0.2718 | 7.0 | 3745 | 0.5440 | 0.5524 | | 0.2532 | 8.0 | 4280 | 0.5431 | 0.5602 | ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

null

Non_BioNLP

# bert-base-uncased-finetuned2-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5176 - Matthews Correlation: 0.5650 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2.6781109393881056e-06 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.5726 | 1.0 | 535 | 0.5090 | 0.3912 | | 0.4467 | 2.0 | 1070 | 0.4536 | 0.5024 | | 0.3891 | 3.0 | 1605 | 0.5093 | 0.4943 | | 0.3387 | 4.0 | 2140 | 0.4927 | 0.5365 | | 0.3177 | 5.0 | 2675 | 0.4897 | 0.5624 | | 0.2853 | 6.0 | 3210 | 0.5176 | 0.5650 | | 0.2718 | 7.0 | 3745 | 0.5440 | 0.5524 | | 0.2532 | 8.0 | 4280 | 0.5431 | 0.5602 | ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-uncased-finetuned2-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "validation", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.5650459791482846, "name": "Matthews Correlation"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,568

BojanSimoski/distilbert-base-uncased-finetuned-cola

BojanSimoski

text-classification

[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-08-17T15:48:29Z

2022-08-17T17:34:52+00:00

10

0

--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: train args: cola metrics: - type: matthews_correlation value: 0.5491398222815213 name: Matthews Correlation ---  # distilbert-base-uncased-finetuned-cola This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5196 - Matthews Correlation: 0.5491 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.5224 | 1.0 | 535 | 0.5262 | 0.4063 | | 0.351 | 2.0 | 1070 | 0.4991 | 0.4871 | | 0.2369 | 3.0 | 1605 | 0.5196 | 0.5491 | | 0.1756 | 4.0 | 2140 | 0.7817 | 0.5142 | | 0.1268 | 5.0 | 2675 | 0.8089 | 0.5324 | ### Framework versions - Transformers 4.21.1 - Pytorch 1.12.1+cu113 - Datasets 2.4.0 - Tokenizers 0.12.1

null

Non_BioNLP

# distilbert-base-uncased-finetuned-cola This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.5196 - Matthews Correlation: 0.5491 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.5224 | 1.0 | 535 | 0.5262 | 0.4063 | | 0.351 | 2.0 | 1070 | 0.4991 | 0.4871 | | 0.2369 | 3.0 | 1605 | 0.5196 | 0.5491 | | 0.1756 | 4.0 | 2140 | 0.7817 | 0.5142 | | 0.1268 | 5.0 | 2675 | 0.8089 | 0.5324 | ### Framework versions - Transformers 4.21.1 - Pytorch 1.12.1+cu113 - Datasets 2.4.0 - Tokenizers 0.12.1

{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "train", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.5491398222815213, "name": "Matthews Correlation"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,569

coverquick/resume-header-classifier

coverquick

text-classification

[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

2023-04-19T07:11:16Z

2023-04-19T07:17:55+00:00

18

0

--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- # resume-header-classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("resume-header-classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

# resume-header-classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("resume-header-classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}

task

[ "TEXT_CLASSIFICATION" ]

44,570

lmstudio-community/internlm2-math-plus-20b-GGUF

lmstudio-community

text-generation

[ "gguf", "math", "text-generation", "en", "zh", "base_model:internlm/internlm2-math-plus-20b", "base_model:quantized:internlm/internlm2-math-plus-20b", "license:other", "endpoints_compatible", "region:us", "imatrix", "conversational" ]

2024-05-30T17:36:00Z

2024-05-30T17:52:27+00:00

317

1

--- base_model: internlm/internlm2-math-plus-20b language: - en - zh license: other pipeline_tag: text-generation tags: - math quantized_by: bartowski lm_studio: param_count: 20b use_case: math release_date: 24-05-2024 model_creator: InternLM prompt_template: ChatML system_prompt: none base_model: InternLM original_repo: internlm/internlm2-math-plus-20b --- ## 💫 Community Model> InternLM2 Math Plus 20b by InternLM *👾 [LM Studio](https://lmstudio.ai) Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on [Discord](https://discord.gg/aPQfnNkxGC)*. **Model creator:** [InternLM](https://huggingface.co/internlm) **Original model**: [internlm2-math-plus-20b](https://huggingface.co/internlm/internlm2-math-plus-20b) **GGUF quantization:** provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b3001](https://github.com/ggerganov/llama.cpp/releases/tag/b3001) ## Model Summary: InternLM2 Math Plus is a series of math proficient models by InternLM, following up on their original series of math models. This series has state of the art bilingual open-sourced math reasoning models at several sizes. This should be used as a solver, prover, verifier, augmentor, with chain of thought reasoning. ## Prompt template: Choose the `ChatML` preset in your LM Studio. Under the hood, the model will see a prompt that's formatted like so: ``` <|im_start|>user {prompt}<|im_end|> <|im_start|>assistant ``` ## Technical Details Math Plus has improved informal math reasoning performance (chain-of-thought and code-intepreter) and formal math reasoning performance (LEAN 4 translation and LEAN 4 theorem proving). InternLM2-Math are continued pretrained from InternLM2-Base with ~100B high quality math-related tokens and SFT with ~2M bilingual math supervised data. More details can be found here: https://github.com/InternLM/InternLM-Math ## Special thanks 🙏 Special thanks to [Georgi Gerganov](https://github.com/ggerganov) and the whole team working on [llama.cpp](https://github.com/ggerganov/llama.cpp/) 🙏 Special thanks to [Kalomaze](https://github.com/kalomaze) and [Dampf](https://github.com/Dampfinchen) for their work on the dataset (linked [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)) that was used for calculating the imatrix for all sizes. ## Disclaimers LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model. You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models. LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.

null

Non_BioNLP

## 💫 Community Model> InternLM2 Math Plus 20b by InternLM *👾 [LM Studio](https://lmstudio.ai) Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on [Discord](https://discord.gg/aPQfnNkxGC)*. **Model creator:** [InternLM](https://huggingface.co/internlm) **Original model**: [internlm2-math-plus-20b](https://huggingface.co/internlm/internlm2-math-plus-20b) **GGUF quantization:** provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b3001](https://github.com/ggerganov/llama.cpp/releases/tag/b3001) ## Model Summary: InternLM2 Math Plus is a series of math proficient models by InternLM, following up on their original series of math models. This series has state of the art bilingual open-sourced math reasoning models at several sizes. This should be used as a solver, prover, verifier, augmentor, with chain of thought reasoning. ## Prompt template: Choose the `ChatML` preset in your LM Studio. Under the hood, the model will see a prompt that's formatted like so: ``` <|im_start|>user {prompt}<|im_end|> <|im_start|>assistant ``` ## Technical Details Math Plus has improved informal math reasoning performance (chain-of-thought and code-intepreter) and formal math reasoning performance (LEAN 4 translation and LEAN 4 theorem proving). InternLM2-Math are continued pretrained from InternLM2-Base with ~100B high quality math-related tokens and SFT with ~2M bilingual math supervised data. More details can be found here: https://github.com/InternLM/InternLM-Math ## Special thanks 🙏 Special thanks to [Georgi Gerganov](https://github.com/ggerganov) and the whole team working on [llama.cpp](https://github.com/ggerganov/llama.cpp/) 🙏 Special thanks to [Kalomaze](https://github.com/kalomaze) and [Dampf](https://github.com/Dampfinchen) for their work on the dataset (linked [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)) that was used for calculating the imatrix for all sizes. ## Disclaimers LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model. You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models. LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.

{"base_model": "internlm/internlm2-math-plus-20b", "language": ["en", "zh"], "license": "other", "pipeline_tag": "text-generation", "tags": ["math"], "quantized_by": "bartowski", "lm_studio": {"param_count": "20b", "use_case": "math", "release_date": "24-05-2024", "model_creator": "InternLM", "prompt_template": "ChatML", "system_prompt": "none", "base_model": "InternLM", "original_repo": "internlm/internlm2-math-plus-20b"}}

task

[ "TRANSLATION" ]

44,571

dimcall/distilbert-base-uncased-finetuned-emotion

dimcall

text-classification

[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:emotion", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-10-02T14:55:47Z

2023-10-02T15:06:24+00:00

7

0

--- datasets: - emotion license: apache-2.0 metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion args: split metrics: - type: accuracy value: 0.9265 name: Accuracy - type: f1 value: 0.9265776722587405 name: F1 ---  # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2129 - Accuracy: 0.9265 - F1: 0.9266 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 0.8255 | 1.0 | 250 | 0.3174 | 0.901 | 0.8980 | | 0.2503 | 2.0 | 500 | 0.2129 | 0.9265 | 0.9266 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.11.0+cu115 - Datasets 1.16.1 - Tokenizers 0.14.0

null

Non_BioNLP

# distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2129 - Accuracy: 0.9265 - F1: 0.9266 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 0.8255 | 1.0 | 250 | 0.3174 | 0.901 | 0.8980 | | 0.2503 | 2.0 | 500 | 0.2129 | 0.9265 | 0.9266 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.11.0+cu115 - Datasets 1.16.1 - Tokenizers 0.14.0

{"datasets": ["emotion"], "license": "apache-2.0", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-emotion", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "args": "split"}, "metrics": [{"type": "accuracy", "value": 0.9265, "name": "Accuracy"}, {"type": "f1", "value": 0.9265776722587405, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,572

TheBloke/medicine-LLM-AWQ

TheBloke

text-generation

[ "transformers", "safetensors", "llama", "text-generation", "biology", "medical", "en", "dataset:Open-Orca/OpenOrca", "dataset:GAIR/lima", "dataset:WizardLM/WizardLM_evol_instruct_V2_196k", "dataset:EleutherAI/pile", "arxiv:2309.09530", "base_model:AdaptLLM/medicine-LLM", "base_model:quantized:AdaptLLM/medicine-LLM", "license:other", "autotrain_compatible", "text-generation-inference", "4-bit", "awq", "region:us" ]

2024-01-15T22:38:37Z

2024-01-15T22:54:43+00:00

100

3

--- base_model: AdaptLLM/medicine-LLM datasets: - Open-Orca/OpenOrca - GAIR/lima - WizardLM/WizardLM_evol_instruct_V2_196k - EleutherAI/pile language: - en license: other metrics: - accuracy model_name: Medicine LLM pipeline_tag: text-generation tags: - biology - medical inference: false model_creator: AdaptLLM model_type: llama prompt_template: '### User Input: {prompt} ### Assistant Output: ' quantized_by: TheBloke ---    <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <div style="display: flex; justify-content: space-between; width: 100%;"> <div style="display: flex; flex-direction: column; align-items: flex-start;"> <a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a> </div> <div style="display: flex; flex-direction: column; align-items: flex-end;"> <a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a> </div> </div> <div style="text-align:center; margin-top: 0em; margin-bottom: 0em">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">  # Medicine LLM - AWQ - Model creator: [AdaptLLM](https://huggingface.co/AdaptLLM) - Original model: [Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM)  ## Description This repo contains AWQ model files for [AdaptLLM's Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM). These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/). ### About AWQ AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings. AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead. It is supported by: - [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ - [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code   ## Repositories available * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/medicine-LLM-AWQ) * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/medicine-LLM-GPTQ) * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/medicine-LLM-GGUF) * [AdaptLLM's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/AdaptLLM/medicine-LLM)   ## Prompt template: AdaptLLM ``` ### User Input: {prompt} ### Assistant Output: ```   ## Provided files, and AWQ parameters I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered. Models are released as sharded safetensors files. | Branch | Bits | GS | AWQ Dataset | Seq Len | Size | | ------ | ---- | -- | ----------- | ------- | ---- | | [main](https://huggingface.co/TheBloke/medicine-LLM-AWQ/tree/main) | 4 | 128 | [Medical Medaow WikiDoc](https://huggingface.co/datasets/medalpaca/medical_meadow_wikidoc/viewer/) | 2048 | 3.89 GB   ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui) Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui). It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install. 1. Click the **Model tab**. 2. Under **Download custom model or LoRA**, enter `TheBloke/medicine-LLM-AWQ`. 3. Click **Download**. 4. The model will start downloading. Once it's finished it will say "Done". 5. In the top left, click the refresh icon next to **Model**. 6. In the **Model** dropdown, choose the model you just downloaded: `medicine-LLM-AWQ` 7. Select **Loader: AutoAWQ**. 8. Click Load, and the model will load and is now ready for use. 9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right. 10. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!   ## Multi-user inference server: vLLM Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/). - Please ensure you are using vLLM version 0.2 or later. - When using vLLM as a server, pass the `--quantization awq` parameter. For example: ```shell python3 -m vllm.entrypoints.api_server --model TheBloke/medicine-LLM-AWQ --quantization awq --dtype auto ``` - When using vLLM from Python code, again set `quantization=awq`. For example: ```python from vllm import LLM, SamplingParams prompts = [ "Tell me about AI", "Write a story about llamas", "What is 291 - 150?", "How much wood would a woodchuck chuck if a woodchuck could chuck wood?", ] prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' prompts = [prompt_template.format(prompt=prompt) for prompt in prompts] sampling_params = SamplingParams(temperature=0.8, top_p=0.95) llm = LLM(model="TheBloke/medicine-LLM-AWQ", quantization="awq", dtype="auto") outputs = llm.generate(prompts, sampling_params) # Print the outputs. for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ```   ## Multi-user inference server: Hugging Face Text Generation Inference (TGI) Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0` Example Docker parameters: ```shell --model-id TheBloke/medicine-LLM-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096 ``` Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later): ```shell pip3 install huggingface-hub ``` ```python from huggingface_hub import InferenceClient endpoint_url = "https://your-endpoint-url-here" prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' client = InferenceClient(endpoint_url) response = client.text_generation(prompt, max_new_tokens=128, do_sample=True, temperature=0.7, top_p=0.95, top_k=40, repetition_penalty=1.1) print(f"Model output: ", response) ```   ## Inference from Python code using Transformers ### Install the necessary packages - Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later. - Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later. ```shell pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0" ``` Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0. If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command: ```shell pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl ``` If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead: ```shell pip3 uninstall -y autoawq git clone https://github.com/casper-hansen/AutoAWQ cd AutoAWQ pip3 install . ``` ### Transformers example code (requires Transformers 4.35.0 and later) ```python from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer model_name_or_path = "TheBloke/medicine-LLM-AWQ" tokenizer = AutoTokenizer.from_pretrained(model_name_or_path) model = AutoModelForCausalLM.from_pretrained( model_name_or_path, low_cpu_mem_usage=True, device_map="cuda:0" ) # Using the text streamer to stream output one token at a time streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True) prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' # Convert prompt to tokens tokens = tokenizer( prompt_template, return_tensors='pt' ).input_ids.cuda() generation_params = { "do_sample": True, "temperature": 0.7, "top_p": 0.95, "top_k": 40, "max_new_tokens": 512, "repetition_penalty": 1.1 } # Generate streamed output, visible one token at a time generation_output = model.generate( tokens, streamer=streamer, **generation_params ) # Generation without a streamer, which will include the prompt in the output generation_output = model.generate( tokens, **generation_params ) # Get the tokens from the output, decode them, print them token_output = generation_output[0] text_output = tokenizer.decode(token_output) print("model.generate output: ", text_output) # Inference is also possible via Transformers' pipeline from transformers import pipeline pipe = pipeline( "text-generation", model=model, tokenizer=tokenizer, **generation_params ) pipe_output = pipe(prompt_template)[0]['generated_text'] print("pipeline output: ", pipe_output) ```   ## Compatibility The files provided are tested to work with: - [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`. - [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later. - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later. - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later.    ## Discord For further support, and discussions on these models and AI in general, join us at: [TheBloke AI's Discord server](https://discord.gg/theblokeai) ## Thanks, and how to contribute Thanks to the [chirper.ai](https://chirper.ai) team! Thanks to Clay from [gpus.llm-utils.org](llm-utils)! I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training. If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects. Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits. * Patreon: https://patreon.com/TheBlokeAI * Ko-Fi: https://ko-fi.com/TheBlokeAI **Special thanks to**: Aemon Algiz. **Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros Thank you to all my generous patrons and donaters! And thank you again to a16z for their generous grant.  # Original model card: AdaptLLM's Medicine LLM # Adapt (Large) Language Models to Domains This repo contains the domain-specific base model developed from **LLaMA-1-7B**, using the method in our paper [Adapting Large Language Models via Reading Comprehension](https://huggingface.co/papers/2309.09530). We explore **continued pre-training on domain-specific corpora** for large language models. While this approach enriches LLMs with domain knowledge, it significantly hurts their prompting ability for question answering. Inspired by human learning via reading comprehension, we propose a simple method to **transform large-scale pre-training corpora into reading comprehension texts**, consistently improving prompting performance across tasks in biomedicine, finance, and law domains. **Our 7B model competes with much larger domain-specific models like BloombergGPT-50B**. ### 🤗 We are currently working hard on developing models across different domains, scales and architectures! Please stay tuned! 🤗 **************************** **Updates** **************************** * 12/19: Released our [13B base models](https://huggingface.co/AdaptLLM/medicine-LLM-13B) developed from LLaMA-1-13B. * 12/8: Released our [chat models](https://huggingface.co/AdaptLLM/medicine-chat) developed from LLaMA-2-Chat-7B. * 9/18: Released our [paper](https://huggingface.co/papers/2309.09530), [code](https://github.com/microsoft/LMOps), [data](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), and [base models](https://huggingface.co/AdaptLLM/medicine-LLM) developed from LLaMA-1-7B. ## Domain-Specific LLaMA-1 ### LLaMA-1-7B In our paper, we develop three domain-specific models from LLaMA-1-7B, which are also available in Huggingface: [Biomedicine-LLM](https://huggingface.co/AdaptLLM/medicine-LLM), [Finance-LLM](https://huggingface.co/AdaptLLM/finance-LLM) and [Law-LLM](https://huggingface.co/AdaptLLM/law-LLM), the performances of our AdaptLLM compared to other domain-specific LLMs are: <img src="https://hf.fast360.xyz/production/uploads/650801ced5578ef7e20b33d4/6efPwitFgy-pLTzvccdcP.png" width="700"> ### LLaMA-1-13B Moreover, we scale up our base model to LLaMA-1-13B to see if **our method is similarly effective for larger-scale models**, and the results are consistently positive too: [Biomedicine-LLM-13B](https://huggingface.co/AdaptLLM/medicine-LLM-13B), [Finance-LLM-13B](https://huggingface.co/AdaptLLM/finance-LLM-13B) and [Law-LLM-13B](https://huggingface.co/AdaptLLM/law-LLM-13B). ## Domain-Specific LLaMA-2-Chat Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our **reading comprehension can perfectly fit the data format** by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat) For example, to chat with the biomedicine base model (**🤗we highly recommend switching to the [chat model](https://huggingface.co/AdaptLLM/medicine-chat) for better response quality!**): ```python from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained("AdaptLLM/medicine-LLM") tokenizer = AutoTokenizer.from_pretrained("AdaptLLM/medicine-LLM", use_fast=False) # Put your input here: user_input = '''Question: Which of the following is an example of monosomy? Options: - 46,XX - 47,XXX - 69,XYY - 45,X Please provide your choice first and then provide explanations if possible.''' # Simply use your input as the prompt for base models prompt = user_input inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device) outputs = model.generate(input_ids=inputs, max_length=2048)[0] answer_start = int(inputs.shape[-1]) pred = tokenizer.decode(outputs[answer_start:], skip_special_tokens=True) print(f'### User Input:\n{user_input}\n\n### Assistant Output:\n{pred}') ``` ## Domain-Specific Tasks To easily reproduce our results, we have uploaded the filled-in zero/few-shot input instructions and output completions of each domain-specific task: [biomedicine-tasks](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), [finance-tasks](https://huggingface.co/datasets/AdaptLLM/finance-tasks), and [law-tasks](https://huggingface.co/datasets/AdaptLLM/law-tasks). **Note:** those filled-in instructions are specifically tailored for models before alignment and do NOT fit for the specific data format required for chat models. ## Citation If you find our work helpful, please cite us: ```bibtex @article{adaptllm, title = {Adapting Large Language Models via Reading Comprehension}, author = {Daixuan Cheng and Shaohan Huang and Furu Wei}, journal = {CoRR}, volume = {abs/2309.09530}, year = {2023} } ```

null

BioNLP

<div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <div style="display: flex; justify-content: space-between; width: 100%;"> <div style="display: flex; flex-direction: column; align-items: flex-start;"> <a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a> </div> <div style="display: flex; flex-direction: column; align-items: flex-end;"> <a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a> </div> </div> <div style="text-align:center; margin-top: 0em; margin-bottom: 0em">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">  # Medicine LLM - AWQ - Model creator: [AdaptLLM](https://huggingface.co/AdaptLLM) - Original model: [Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM)  ## Description This repo contains AWQ model files for [AdaptLLM's Medicine LLM](https://huggingface.co/AdaptLLM/medicine-LLM). These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/). ### About AWQ AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference with equivalent or better quality compared to the most commonly used GPTQ settings. AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead. It is supported by: - [Text Generation Webui](https://github.com/oobabooga/text-generation-webui) - using Loader: AutoAWQ - [vLLM](https://github.com/vllm-project/vllm) - version 0.2.2 or later for support for all model types. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later, from any code or client that supports Transformers - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) - for use from Python code   ## Repositories available * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/medicine-LLM-AWQ) * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/medicine-LLM-GPTQ) * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/medicine-LLM-GGUF) * [AdaptLLM's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/AdaptLLM/medicine-LLM)   ## Prompt template: AdaptLLM ``` ### User Input: {prompt} ### Assistant Output: ```   ## Provided files, and AWQ parameters I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered. Models are released as sharded safetensors files. | Branch | Bits | GS | AWQ Dataset | Seq Len | Size | | ------ | ---- | -- | ----------- | ------- | ---- | | [main](https://huggingface.co/TheBloke/medicine-LLM-AWQ/tree/main) | 4 | 128 | [Medical Medaow WikiDoc](https://huggingface.co/datasets/medalpaca/medical_meadow_wikidoc/viewer/) | 2048 | 3.89 GB   ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui) Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui). It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install. 1. Click the **Model tab**. 2. Under **Download custom model or LoRA**, enter `TheBloke/medicine-LLM-AWQ`. 3. Click **Download**. 4. The model will start downloading. Once it's finished it will say "Done". 5. In the top left, click the refresh icon next to **Model**. 6. In the **Model** dropdown, choose the model you just downloaded: `medicine-LLM-AWQ` 7. Select **Loader: AutoAWQ**. 8. Click Load, and the model will load and is now ready for use. 9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right. 10. Once you're ready, click the **Text Generation** tab and enter a prompt to get started!   ## Multi-user inference server: vLLM Documentation on installing and using vLLM [can be found here](https://vllm.readthedocs.io/en/latest/). - Please ensure you are using vLLM version 0.2 or later. - When using vLLM as a server, pass the `--quantization awq` parameter. For example: ```shell python3 -m vllm.entrypoints.api_server --model TheBloke/medicine-LLM-AWQ --quantization awq --dtype auto ``` - When using vLLM from Python code, again set `quantization=awq`. For example: ```python from vllm import LLM, SamplingParams prompts = [ "Tell me about AI", "Write a story about llamas", "What is 291 - 150?", "How much wood would a woodchuck chuck if a woodchuck could chuck wood?", ] prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' prompts = [prompt_template.format(prompt=prompt) for prompt in prompts] sampling_params = SamplingParams(temperature=0.8, top_p=0.95) llm = LLM(model="TheBloke/medicine-LLM-AWQ", quantization="awq", dtype="auto") outputs = llm.generate(prompts, sampling_params) # Print the outputs. for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ```   ## Multi-user inference server: Hugging Face Text Generation Inference (TGI) Use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0` Example Docker parameters: ```shell --model-id TheBloke/medicine-LLM-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096 ``` Example Python code for interfacing with TGI (requires [huggingface-hub](https://github.com/huggingface/huggingface_hub) 0.17.0 or later): ```shell pip3 install huggingface-hub ``` ```python from huggingface_hub import InferenceClient endpoint_url = "https://your-endpoint-url-here" prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' client = InferenceClient(endpoint_url) response = client.text_generation(prompt, max_new_tokens=128, do_sample=True, temperature=0.7, top_p=0.95, top_k=40, repetition_penalty=1.1) print(f"Model output: ", response) ```   ## Inference from Python code using Transformers ### Install the necessary packages - Requires: [Transformers](https://huggingface.co/docs/transformers) 4.35.0 or later. - Requires: [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) 0.1.6 or later. ```shell pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0" ``` Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0. If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command: ```shell pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl ``` If you have problems installing [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) using the pre-built wheels, install it from source instead: ```shell pip3 uninstall -y autoawq git clone https://github.com/casper-hansen/AutoAWQ cd AutoAWQ pip3 install . ``` ### Transformers example code (requires Transformers 4.35.0 and later) ```python from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer model_name_or_path = "TheBloke/medicine-LLM-AWQ" tokenizer = AutoTokenizer.from_pretrained(model_name_or_path) model = AutoModelForCausalLM.from_pretrained( model_name_or_path, low_cpu_mem_usage=True, device_map="cuda:0" ) # Using the text streamer to stream output one token at a time streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True) prompt = "Tell me about AI" prompt_template=f'''### User Input: {prompt} ### Assistant Output: ''' # Convert prompt to tokens tokens = tokenizer( prompt_template, return_tensors='pt' ).input_ids.cuda() generation_params = { "do_sample": True, "temperature": 0.7, "top_p": 0.95, "top_k": 40, "max_new_tokens": 512, "repetition_penalty": 1.1 } # Generate streamed output, visible one token at a time generation_output = model.generate( tokens, streamer=streamer, **generation_params ) # Generation without a streamer, which will include the prompt in the output generation_output = model.generate( tokens, **generation_params ) # Get the tokens from the output, decode them, print them token_output = generation_output[0] text_output = tokenizer.decode(token_output) print("model.generate output: ", text_output) # Inference is also possible via Transformers' pipeline from transformers import pipeline pipe = pipeline( "text-generation", model=model, tokenizer=tokenizer, **generation_params ) pipe_output = pipe(prompt_template)[0]['generated_text'] print("pipeline output: ", pipe_output) ```   ## Compatibility The files provided are tested to work with: - [text-generation-webui](https://github.com/oobabooga/text-generation-webui) using `Loader: AutoAWQ`. - [vLLM](https://github.com/vllm-project/vllm) version 0.2.0 and later. - [Hugging Face Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) version 1.1.0 and later. - [Transformers](https://huggingface.co/docs/transformers) version 4.35.0 and later. - [AutoAWQ](https://github.com/casper-hansen/AutoAWQ) version 0.1.1 and later.    ## Discord For further support, and discussions on these models and AI in general, join us at: [TheBloke AI's Discord server](https://discord.gg/theblokeai) ## Thanks, and how to contribute Thanks to the [chirper.ai](https://chirper.ai) team! Thanks to Clay from [gpus.llm-utils.org](llm-utils)! I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training. If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects. Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits. * Patreon: https://patreon.com/TheBlokeAI * Ko-Fi: https://ko-fi.com/TheBlokeAI **Special thanks to**: Aemon Algiz. **Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros Thank you to all my generous patrons and donaters! And thank you again to a16z for their generous grant.  # Original model card: AdaptLLM's Medicine LLM # Adapt (Large) Language Models to Domains This repo contains the domain-specific base model developed from **LLaMA-1-7B**, using the method in our paper [Adapting Large Language Models via Reading Comprehension](https://huggingface.co/papers/2309.09530). We explore **continued pre-training on domain-specific corpora** for large language models. While this approach enriches LLMs with domain knowledge, it significantly hurts their prompting ability for question answering. Inspired by human learning via reading comprehension, we propose a simple method to **transform large-scale pre-training corpora into reading comprehension texts**, consistently improving prompting performance across tasks in biomedicine, finance, and law domains. **Our 7B model competes with much larger domain-specific models like BloombergGPT-50B**. ### 🤗 We are currently working hard on developing models across different domains, scales and architectures! Please stay tuned! 🤗 **************************** **Updates** **************************** * 12/19: Released our [13B base models](https://huggingface.co/AdaptLLM/medicine-LLM-13B) developed from LLaMA-1-13B. * 12/8: Released our [chat models](https://huggingface.co/AdaptLLM/medicine-chat) developed from LLaMA-2-Chat-7B. * 9/18: Released our [paper](https://huggingface.co/papers/2309.09530), [code](https://github.com/microsoft/LMOps), [data](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), and [base models](https://huggingface.co/AdaptLLM/medicine-LLM) developed from LLaMA-1-7B. ## Domain-Specific LLaMA-1 ### LLaMA-1-7B In our paper, we develop three domain-specific models from LLaMA-1-7B, which are also available in Huggingface: [Biomedicine-LLM](https://huggingface.co/AdaptLLM/medicine-LLM), [Finance-LLM](https://huggingface.co/AdaptLLM/finance-LLM) and [Law-LLM](https://huggingface.co/AdaptLLM/law-LLM), the performances of our AdaptLLM compared to other domain-specific LLMs are: <img src="https://hf.fast360.xyz/production/uploads/650801ced5578ef7e20b33d4/6efPwitFgy-pLTzvccdcP.png" width="700"> ### LLaMA-1-13B Moreover, we scale up our base model to LLaMA-1-13B to see if **our method is similarly effective for larger-scale models**, and the results are consistently positive too: [Biomedicine-LLM-13B](https://huggingface.co/AdaptLLM/medicine-LLM-13B), [Finance-LLM-13B](https://huggingface.co/AdaptLLM/finance-LLM-13B) and [Law-LLM-13B](https://huggingface.co/AdaptLLM/law-LLM-13B). ## Domain-Specific LLaMA-2-Chat Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our **reading comprehension can perfectly fit the data format** by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat) For example, to chat with the biomedicine base model (**🤗we highly recommend switching to the [chat model](https://huggingface.co/AdaptLLM/medicine-chat) for better response quality!**): ```python from transformers import AutoModelForCausalLM, AutoTokenizer model = AutoModelForCausalLM.from_pretrained("AdaptLLM/medicine-LLM") tokenizer = AutoTokenizer.from_pretrained("AdaptLLM/medicine-LLM", use_fast=False) # Put your input here: user_input = '''Question: Which of the following is an example of monosomy? Options: - 46,XX - 47,XXX - 69,XYY - 45,X Please provide your choice first and then provide explanations if possible.''' # Simply use your input as the prompt for base models prompt = user_input inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device) outputs = model.generate(input_ids=inputs, max_length=2048)[0] answer_start = int(inputs.shape[-1]) pred = tokenizer.decode(outputs[answer_start:], skip_special_tokens=True) print(f'### User Input:\n{user_input}\n\n### Assistant Output:\n{pred}') ``` ## Domain-Specific Tasks To easily reproduce our results, we have uploaded the filled-in zero/few-shot input instructions and output completions of each domain-specific task: [biomedicine-tasks](https://huggingface.co/datasets/AdaptLLM/medicine-tasks), [finance-tasks](https://huggingface.co/datasets/AdaptLLM/finance-tasks), and [law-tasks](https://huggingface.co/datasets/AdaptLLM/law-tasks). **Note:** those filled-in instructions are specifically tailored for models before alignment and do NOT fit for the specific data format required for chat models. ## Citation If you find our work helpful, please cite us: ```bibtex @article{adaptllm, title = {Adapting Large Language Models via Reading Comprehension}, author = {Daixuan Cheng and Shaohan Huang and Furu Wei}, journal = {CoRR}, volume = {abs/2309.09530}, year = {2023} } ```

{"base_model": "AdaptLLM/medicine-LLM", "datasets": ["Open-Orca/OpenOrca", "GAIR/lima", "WizardLM/WizardLM_evol_instruct_V2_196k", "EleutherAI/pile"], "language": ["en"], "license": "other", "metrics": ["accuracy"], "model_name": "Medicine LLM", "pipeline_tag": "text-generation", "tags": ["biology", "medical"], "inference": false, "model_creator": "AdaptLLM", "model_type": "llama", "prompt_template": "### User Input:\n{prompt}\n\n### Assistant Output:\n", "quantized_by": "TheBloke"}

task

[ "QUESTION_ANSWERING" ]

44,573

zenml/finetuned-snowflake-arctic-embed-m

zenml

sentence-similarity

[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:1490", "loss:MatryoshkaLoss", "loss:TripletLoss", "en", "arxiv:1908.10084", "arxiv:2205.13147", "arxiv:1703.07737", "base_model:Snowflake/snowflake-arctic-embed-m", "base_model:finetune:Snowflake/snowflake-arctic-embed-m", "license:apache-2.0", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

2024-08-05T19:08:54Z

2024-08-08T14:19:44+00:00

25

1

--- base_model: Snowflake/snowflake-arctic-embed-m datasets: [] language: - en library_name: sentence-transformers license: apache-2.0 metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:1490 - loss:MatryoshkaLoss - loss:TripletLoss widget: - source_sentence: Where is the global configuration directory located in ZenML's default setup? sentences: - '''default'' ... Creating default user ''default'' ...Creating default stack for user ''default'' in workspace default... Active workspace not set. Setting it to the default. The active stack is not set. Setting the active stack to the default workspace stack. Using the default store for the global config. Unable to find ZenML repository in your current working directory (/tmp/folder) or any parent directories. If you want to use an existing repository which is in a different location, set the environment variable ''ZENML_REPOSITORY_PATH''. If you want to create a new repository, run zenml init. Running without an active repository root. Using the default local database. Running with active workspace: ''default'' (global) ┏━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━┓ ┃ ACTIVE │ STACK NAME │ SHARED │ OWNER │ ARTIFACT_STORE │ ORCHESTRATOR ┃ ┠────────┼────────────┼────────┼─────────┼────────────────┼──────────────┨ ┃ 👉 │ default │ ❌ │ default │ default │ default ┃ ┗━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━┛ The following is an example of the layout of the global config directory immediately after initialization: /home/stefan/.config/zenml <- Global Config Directory ├── config.yaml <- Global Configuration Settings └── local_stores <- Every Stack component that stores information | locally will have its own subdirectory here. ├── a1a0d3d0-d552-4a80-be09-67e5e29be8ee <- e.g. Local Store path for the | `default` local Artifact Store └── default_zen_store └── zenml.db <- SQLite database where ZenML data (stacks, components, etc) are stored by default. As shown above, the global config directory stores the following information:' - How do you configure the network settings on a Linux server? - 'Reranking for better retrieval Add reranking to your RAG inference for better retrieval performance. Rerankers are a crucial component of retrieval systems that use LLMs. They help improve the quality of the retrieved documents by reordering them based on additional features or scores. In this section, we''ll explore how to add a reranker to your RAG inference pipeline in ZenML. In previous sections, we set up the overall workflow, from data ingestion and preprocessing to embeddings generation and retrieval. We then set up some basic evaluation metrics to assess the performance of our retrieval system. A reranker is a way to squeeze a bit of extra performance out of the system by reordering the retrieved documents based on additional features or scores. As you can see, reranking is an optional addition we make to what we''ve already set up. It''s not strictly necessary, but it can help improve the relevance and quality of the retrieved documents, which in turn can lead to better responses from the LLM. Let''s dive in! PreviousEvaluation in practice NextUnderstanding reranking Last updated 1 month ago' - source_sentence: Where can I find the instructions to enable CUDA for GPU-backed hardware in ZenML SDK Docs? sentences: - 'Migration guide 0.39.1 → 0.41.0 How to migrate your ZenML pipelines and steps from version <=0.39.1 to 0.41.0. ZenML versions 0.40.0 to 0.41.0 introduced a new and more flexible syntax to define ZenML steps and pipelines. This page contains code samples that show you how to upgrade your steps and pipelines to the new syntax. Newer versions of ZenML still work with pipelines and steps defined using the old syntax, but the old syntax is deprecated and will be removed in the future. Overview from typing import Optional from zenml.steps import BaseParameters, Output, StepContext, step from zenml.pipelines import pipeline # Define a Step class MyStepParameters(BaseParameters): param_1: int param_2: Optional[float] = None @step def my_step( params: MyStepParameters, context: StepContext, ) -> Output(int_output=int, str_output=str): result = int(params.param_1 * (params.param_2 or 1)) result_uri = context.get_output_artifact_uri() return result, result_uri # Run the Step separately my_step.entrypoint() # Define a Pipeline @pipeline def my_pipeline(my_step): my_step() step_instance = my_step(params=MyStepParameters(param_1=17)) pipeline_instance = my_pipeline(my_step=step_instance) # Configure and run the Pipeline pipeline_instance.configure(enable_cache=False) schedule = Schedule(...) pipeline_instance.run(schedule=schedule) # Fetch the Pipeline Run last_run = pipeline_instance.get_runs()[0] int_output = last_run.get_step["my_step"].outputs["int_output"].read() from typing import Annotated, Optional, Tuple from zenml import get_step_context, pipeline, step from zenml.client import Client # Define a Step @step def my_step( param_1: int, param_2: Optional[float] = None ) -> Tuple[Annotated[int, "int_output"], Annotated[str, "str_output"]]: result = int(param_1 * (param_2 or 1)) result_uri = get_step_context().get_output_artifact_uri() return result, result_uri # Run the Step separately my_step() # Define a Pipeline @pipeline' - How do I integrate Google Cloud VertexAI into my existing Kubernetes cluster? - ' SDK Docs . Enabling CUDA for GPU-backed hardwareNote that if you wish to use this step operator to run steps on a GPU, you will need to follow the instructions on this page to ensure that it works. It requires adding some extra settings customization and is essential to enable CUDA for the GPU to give its full acceleration. PreviousStep Operators NextGoogle Cloud VertexAI Last updated 19 days ago' - source_sentence: What are the special metadata types supported by ZenML and how are they used? sentences: - 'Special Metadata Types Tracking your metadata. ZenML supports several special metadata types to capture specific kinds of information. Here are examples of how to use the special types Uri, Path, DType, and StorageSize: from zenml.metadata.metadata_types import StorageSize, DType from zenml import log_artifact_metadata log_artifact_metadata( metadata={ "dataset_source": Uri("gs://my-bucket/datasets/source.csv"), "preprocessing_script": Path("/scripts/preprocess.py"), "column_types": { "age": DType("int"), "income": DType("float"), "score": DType("int") }, "processed_data_size": StorageSize(2500000) In this example: Uri is used to indicate a dataset source URI. Path is used to specify the filesystem path to a preprocessing script. DType is used to describe the data types of specific columns. StorageSize is used to indicate the size of the processed data in bytes. These special types help standardize the format of metadata and ensure that it is logged in a consistent and interpretable manner. PreviousGroup metadata NextFetch metadata within steps Last updated 19 days ago' - 'Configure a code repository Connect a Git repository to ZenML to track code changes and collaborate on MLOps projects. Throughout the lifecycle of a MLOps pipeline, it can get quite tiresome to always wait for a Docker build every time after running a pipeline (even if the local Docker cache is used). However, there is a way to just have one pipeline build and keep reusing it until a change to the pipeline environment is made: by connecting a code repository. With ZenML, connecting to a Git repository optimizes the Docker build processes. It also has the added bonus of being a better way of managing repository changes and enabling better code collaboration. Here is how the flow changes when running a pipeline: You trigger a pipeline run on your local machine. ZenML parses the @pipeline function to determine the necessary steps. The local client requests stack information from the ZenML server, which responds with the cloud stack configuration. The local client detects that we''re using a code repository and requests the information from the git repo. Instead of building a new Docker image, the client checks if an existing image can be reused based on the current Git commit hash and other environment metadata. The client initiates a run in the orchestrator, which sets up the execution environment in the cloud, such as a VM. The orchestrator downloads the code directly from the Git repository and uses the existing Docker image to run the pipeline steps. Pipeline steps execute, storing artifacts in the cloud-based artifact store. Throughout the execution, the pipeline run status and metadata are reported back to the ZenML server. By connecting a Git repository, you avoid redundant builds and make your MLOps processes more efficient. Your team can work on the codebase simultaneously, with ZenML handling the version tracking and ensuring that the correct code version is always used for each run. Creating a GitHub Repository' - Can you explain the process of setting up a virtual environment in Python? - source_sentence: What are the benefits of deploying stack components directly from the ZenML CLI? sentences: - '─────────────────────────────────────────────────┨┃ RESOURCE TYPES │ 🔵 gcp-generic, 📦 gcs-bucket, 🌀 kubernetes-cluster, 🐳 docker-registry ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ RESOURCE NAME │ <multiple> ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ SECRET ID │ 4694de65-997b-4929-8831-b49d5e067b97 ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ SESSION DURATION │ N/A ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ EXPIRES IN │ 59m46s ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ OWNER │ default ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ WORKSPACE │ default ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ SHARED │ ➖ ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ CREATED_AT │ 2023-05-19 09:04:33.557126 ┃ ┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨ ┃ UPDATED_AT │ 2023-05-19 09:04:33.557127 ┃ ┗━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛ Configuration ┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓' - How do you set up a custom service account for Vertex AI? - '⚒️Manage stacks Deploying your stack components directly from the ZenML CLI The first step in running your pipelines on remote infrastructure is to deploy all the components that you would need, like an MLflow tracking server, a Seldon Core model deployer, and more to your cloud. This can bring plenty of benefits like scalability, reliability, and collaboration. ZenML eases the path to production by providing a seamless way for all tools to interact with others through the use of abstractions. However, one of the most painful parts of this process, from what we see on our Slack and in general, is the deployment of these stack components. Deploying and managing MLOps tools is tricky 😭😵‍💫 It is not trivial to set up all the different tools that you might need for your pipeline. 🌈 Each tool comes with a certain set of requirements. For example, a Kubeflow installation will require you to have a Kubernetes cluster, and so would a Seldon Core deployment. 🤔 Figuring out the defaults for infra parameters is not easy. Even if you have identified the backing infra that you need for a stack component, setting up reasonable defaults for parameters like instance size, CPU, memory, etc., needs a lot of experimentation to figure out. 🚧 Many times, standard tool installations don''t work out of the box. For example, to run a custom pipeline in Vertex AI, it is not enough to just run an imported pipeline. You might also need a custom service account that is configured to perform tasks like reading secrets from your secret store or talking to other GCP services that your pipeline might need. 🔐 Some tools need an additional layer of installations to enable a more secure, production-grade setup. For example, a standard MLflow tracking server deployment comes without an authentication frontend which might expose all of your tracking data to the world if deployed as-is.' - source_sentence: What is the expiration time for the GCP OAuth2 token in the ZenML configuration? sentences: - '━━━━━┛ Configuration ┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃ ┠────────────┼────────────┨ ┃ project_id │ zenml-core ┃ ┠────────────┼────────────┨ ┃ token │ [HIDDEN] ┃ ┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛ Note the temporary nature of the Service Connector. It will expire and become unusable in 1 hour: zenml service-connector list --name gcp-oauth2-token Example Command Output ┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓ ┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃ ┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨ ┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃ ┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃ ┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃ ┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃ ┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛ Auto-configuration The GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.' - 'Hugging Face Deploying models to Huggingface Inference Endpoints with Hugging Face :hugging_face:. Hugging Face Inference Endpoints provides a secure production solution to easily deploy any transformers, sentence-transformers, and diffusers models on a dedicated and autoscaling infrastructure managed by Hugging Face. An Inference Endpoint is built from a model from the Hub. This service provides dedicated and autoscaling infrastructure managed by Hugging Face, allowing you to deploy models without dealing with containers and GPUs. When to use it? You should use Hugging Face Model Deployer: if you want to deploy Transformers, Sentence-Transformers, or Diffusion models on dedicated and secure infrastructure. if you prefer a fully-managed production solution for inference without the need to handle containers and GPUs. if your goal is to turn your models into production-ready APIs with minimal infrastructure or MLOps involvement Cost-effectiveness is crucial, and you want to pay only for the raw compute resources you use. Enterprise security is a priority, and you need to deploy models into secure offline endpoints accessible only via a direct connection to your Virtual Private Cloud (VPCs). If you are looking for a more easy way to deploy your models locally, you can use the MLflow Model Deployer flavor. How to deploy it? The Hugging Face Model Deployer flavor is provided by the Hugging Face ZenML integration, so you need to install it on your local machine to be able to deploy your models. You can do this by running the following command: zenml integration install huggingface -y To register the Hugging Face model deployer with ZenML you need to run the following command: zenml model-deployer register <MODEL_DEPLOYER_NAME> --flavor=huggingface --token=<YOUR_HF_TOKEN> --namespace=<YOUR_HF_NAMESPACE> Here, token parameter is the Hugging Face authentication token. It can be managed through Hugging Face settings.' - Can you list the steps to set up a Docker registry on a Kubernetes cluster? model-index: - name: zenml/finetuned-snowflake-arctic-embed-m results: - task: type: information-retrieval name: Information Retrieval dataset: name: dim 384 type: dim_384 metrics: - type: cosine_accuracy@1 value: 0.29518072289156627 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.5240963855421686 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5843373493975904 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.6867469879518072 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.29518072289156627 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.17469879518072293 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.11686746987951804 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.0686746987951807 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.29518072289156627 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.5240963855421686 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5843373493975904 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.6867469879518072 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.4908042072911187 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.42844234079173843 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.43576329240226386 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 256 type: dim_256 metrics: - type: cosine_accuracy@1 value: 0.25903614457831325 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.5060240963855421 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5783132530120482 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.6445783132530121 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.25903614457831325 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.1686746987951807 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.11566265060240961 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.0644578313253012 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.25903614457831325 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.5060240963855421 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5783132530120482 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.6445783132530121 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.4548319777111225 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.39346194301013593 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.40343211538391555 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 128 type: dim_128 metrics: - type: cosine_accuracy@1 value: 0.2710843373493976 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.46987951807228917 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5662650602409639 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.6144578313253012 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.2710843373493976 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.1566265060240964 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.11325301204819276 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.061445783132530116 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.2710843373493976 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.46987951807228917 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5662650602409639 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.6144578313253012 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.44433019669319024 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.3893574297188756 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.3989315479842741 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 64 type: dim_64 metrics: - type: cosine_accuracy@1 value: 0.21686746987951808 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.42168674698795183 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.5180722891566265 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.5843373493975904 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.21686746987951808 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.14056224899598396 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.10361445783132528 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.05843373493975902 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.21686746987951808 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.42168674698795183 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.5180722891566265 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.5843373493975904 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.39639025659520544 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.3364529546758464 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.34658882510541217 name: Cosine Map@100 --- # zenml/finetuned-snowflake-arctic-embed-m This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 tokens - **Similarity Function:** Cosine Similarity  - **Language:** en - **License:** apache-2.0 ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("zenml/finetuned-snowflake-arctic-embed-m") # Run inference sentences = [ 'What is the expiration time for the GCP OAuth2 token in the ZenML configuration?', '━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃\n\n┠────────────┼────────────┨\n\n┃ project_id │ zenml-core ┃\n\n┠────────────┼────────────┨\n\n┃ token │ [HIDDEN] ┃\n\n┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛\n\nNote the temporary nature of the Service Connector. It will expire and become unusable in 1 hour:\n\nzenml service-connector list --name gcp-oauth2-token\n\nExample Command Output\n\n┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓\n\n┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃\n\n┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨\n\n┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃\n\n┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛\n\nAuto-configuration\n\nThe GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.', 'Can you list the steps to set up a Docker registry on a Kubernetes cluster?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_384` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.2952 | | cosine_accuracy@3 | 0.5241 | | cosine_accuracy@5 | 0.5843 | | cosine_accuracy@10 | 0.6867 | | cosine_precision@1 | 0.2952 | | cosine_precision@3 | 0.1747 | | cosine_precision@5 | 0.1169 | | cosine_precision@10 | 0.0687 | | cosine_recall@1 | 0.2952 | | cosine_recall@3 | 0.5241 | | cosine_recall@5 | 0.5843 | | cosine_recall@10 | 0.6867 | | cosine_ndcg@10 | 0.4908 | | cosine_mrr@10 | 0.4284 | | **cosine_map@100** | **0.4358** | #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.259 | | cosine_accuracy@3 | 0.506 | | cosine_accuracy@5 | 0.5783 | | cosine_accuracy@10 | 0.6446 | | cosine_precision@1 | 0.259 | | cosine_precision@3 | 0.1687 | | cosine_precision@5 | 0.1157 | | cosine_precision@10 | 0.0645 | | cosine_recall@1 | 0.259 | | cosine_recall@3 | 0.506 | | cosine_recall@5 | 0.5783 | | cosine_recall@10 | 0.6446 | | cosine_ndcg@10 | 0.4548 | | cosine_mrr@10 | 0.3935 | | **cosine_map@100** | **0.4034** | #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.2711 | | cosine_accuracy@3 | 0.4699 | | cosine_accuracy@5 | 0.5663 | | cosine_accuracy@10 | 0.6145 | | cosine_precision@1 | 0.2711 | | cosine_precision@3 | 0.1566 | | cosine_precision@5 | 0.1133 | | cosine_precision@10 | 0.0614 | | cosine_recall@1 | 0.2711 | | cosine_recall@3 | 0.4699 | | cosine_recall@5 | 0.5663 | | cosine_recall@10 | 0.6145 | | cosine_ndcg@10 | 0.4443 | | cosine_mrr@10 | 0.3894 | | **cosine_map@100** | **0.3989** | #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.2169 | | cosine_accuracy@3 | 0.4217 | | cosine_accuracy@5 | 0.5181 | | cosine_accuracy@10 | 0.5843 | | cosine_precision@1 | 0.2169 | | cosine_precision@3 | 0.1406 | | cosine_precision@5 | 0.1036 | | cosine_precision@10 | 0.0584 | | cosine_recall@1 | 0.2169 | | cosine_recall@3 | 0.4217 | | cosine_recall@5 | 0.5181 | | cosine_recall@10 | 0.5843 | | cosine_ndcg@10 | 0.3964 | | cosine_mrr@10 | 0.3365 | | **cosine_map@100** | **0.3466** |   ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 1,490 training samples * Columns: <code>positive</code>, <code>anchor</code>, and <code>negative</code> * Approximate statistics based on the first 1000 samples: | | positive | anchor | negative | |:--------|:----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------| | type | string | string | string | | details | <ul><li>min: 9 tokens</li><li>mean: 21.02 tokens</li><li>max: 64 tokens</li></ul> | <ul><li>min: 23 tokens</li><li>mean: 375.16 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 10 tokens</li><li>mean: 17.51 tokens</li><li>max: 31 tokens</li></ul> | * Samples: | positive | anchor | negative | |:-----------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------| | <code>What details can you provide about the mlflow_training_pipeline runs listed in the ZenML documentation?</code> | <code>mlflow_training_pipeline', ┃┃ │ │ │ 'zenml_pipeline_run_uuid': 'a5d4faae-ef70-48f2-9893-6e65d5e51e98', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.005'} ┃ ┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨ ┃ tensorflow-mnist-model │ 2 │ Run #2 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_09_08_467212', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃ ┃ │ │ │ 'zenml_pipeline_run_uuid': '11858dcf-3e47-4b1a-82c5-6fa25ba4e037', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.003'} ┃ ┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨ ┃ tensorflow-mnist-model │ 1 │ Run #1 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_08_52_398499', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃ ┃ │ │ │ 'zenml_pipeline_run_uuid': '29fb22c1-6e0b-4431-9e04-226226506d16', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.001'} ┃</code> | <code>Can you explain how to configure the TensorFlow settings for a different project?</code> | | <code>How do you register a GCP Service Connector that uses account impersonation to access the zenml-bucket-sl GCS bucket?</code> | <code>esource-id zenml-bucket-sl Example Command OutputError: Service connector 'gcp-empty-sa' verification failed: connector authorization failure: failed to fetch GCS bucket zenml-bucket-sl: 403 GET https://storage.googleapis.com/storage/v1/b/zenml-bucket-sl?projection=noAcl&prettyPrint=false: [email protected] does not have storage.buckets.get access to the Google Cloud Storage bucket. Permission 'storage.buckets.get' denied on resource (or it may not exist). Next, we'll register a GCP Service Connector that actually uses account impersonation to access the zenml-bucket-sl GCS bucket and verify that it can actually access the bucket: zenml service-connector register gcp-impersonate-sa --type gcp --auth-method impersonation --service_account_json=@[email protected] --project_id=zenml-core --target_principal=zenml-bucket-sl@zenml-core.iam.gserviceaccount.com --resource-type gcs-bucket --resource-id gs://zenml-bucket-sl Example Command Output Expanding argument value service_account_json to contents of file /home/stefan/aspyre/src/zenml/[email protected]. Successfully registered service connector `gcp-impersonate-sa` with access to the following resources: ┏━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━┓ ┃ RESOURCE TYPE │ RESOURCE NAMES ┃ ┠───────────────┼──────────────────────┨ ┃ 📦 gcs-bucket │ gs://zenml-bucket-sl ┃ ┗━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━┛ External Account (GCP Workload Identity) Use GCP workload identity federation to authenticate to GCP services using AWS IAM credentials, Azure Active Directory credentials or generic OIDC tokens.</code> | <code>What is the process for setting up a ZenML pipeline using AWS IAM credentials?</code> | | <code>Can you explain how data validation helps in detecting data drift and model drift in ZenML pipelines?</code> | <code>of your models at different stages of development.if you have pipelines that regularly ingest new data, you should use data validation to run regular data integrity checks to signal problems before they are propagated downstream. in continuous training pipelines, you should use data validation techniques to compare new training data against a data reference and to compare the performance of newly trained models against previous ones. when you have pipelines that automate batch inference or if you regularly collect data used as input in online inference, you should use data validation to run data drift analyses and detect training-serving skew, data drift and model drift. Data Validator Flavors Data Validator are optional stack components provided by integrations. The following table lists the currently available Data Validators and summarizes their features and the data types and model types that they can be used with in ZenML pipelines: Data Validator Validation Features Data Types Model Types Notes Flavor/Integration Deepchecks data quality data drift model drift model performance tabular: pandas.DataFrame CV: torch.utils.data.dataloader.DataLoader tabular: sklearn.base.ClassifierMixin CV: torch.nn.Module Add Deepchecks data and model validation tests to your pipelines deepchecks Evidently data quality data drift model drift model performance tabular: pandas.DataFrame N/A Use Evidently to generate a variety of data quality and data/model drift reports and visualizations evidently Great Expectations data profiling data quality tabular: pandas.DataFrame N/A Perform data testing, documentation and profiling with Great Expectations great_expectations Whylogs/WhyLabs data drift tabular: pandas.DataFrame N/A Generate data profiles with whylogs and upload them to WhyLabs whylogs If you would like to see the available flavors of Data Validator, you can use the command: zenml data-validator flavor list How to use it</code> | <code>What are the best practices for deploying web applications using Docker and Kubernetes?</code> | * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "TripletLoss", "matryoshka_dims": [ 384, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `gradient_accumulation_steps`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 4 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 16 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 4 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: True - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs | Epoch | Step | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_384_cosine_map@100 | dim_64_cosine_map@100 | |:-------:|:-----:|:----------------------:|:----------------------:|:----------------------:|:---------------------:| | 0.6667 | 1 | 0.3884 | 0.4332 | 0.4464 | 0.3140 | | **2.0** | **3** | **0.4064** | **0.4195** | **0.4431** | **0.3553** | | 2.6667 | 4 | 0.3989 | 0.4034 | 0.4358 | 0.3466 | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.3.1+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### TripletLoss ```bibtex @misc{hermans2017defense, title={In Defense of the Triplet Loss for Person Re-Identification}, author={Alexander Hermans and Lucas Beyer and Bastian Leibe}, year={2017}, eprint={1703.07737}, archivePrefix={arXiv}, primaryClass={cs.CV} } ```

null

Non_BioNLP

# zenml/finetuned-snowflake-arctic-embed-m This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 tokens - **Similarity Function:** Cosine Similarity  - **Language:** en - **License:** apache-2.0 ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("zenml/finetuned-snowflake-arctic-embed-m") # Run inference sentences = [ 'What is the expiration time for the GCP OAuth2 token in the ZenML configuration?', '━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃\n\n┠────────────┼────────────┨\n\n┃ project_id │ zenml-core ┃\n\n┠────────────┼────────────┨\n\n┃ token │ [HIDDEN] ┃\n\n┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛\n\nNote the temporary nature of the Service Connector. It will expire and become unusable in 1 hour:\n\nzenml service-connector list --name gcp-oauth2-token\n\nExample Command Output\n\n┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓\n\n┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃\n\n┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨\n\n┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃\n\n┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛\n\nAuto-configuration\n\nThe GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.', 'Can you list the steps to set up a Docker registry on a Kubernetes cluster?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_384` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.2952 | | cosine_accuracy@3 | 0.5241 | | cosine_accuracy@5 | 0.5843 | | cosine_accuracy@10 | 0.6867 | | cosine_precision@1 | 0.2952 | | cosine_precision@3 | 0.1747 | | cosine_precision@5 | 0.1169 | | cosine_precision@10 | 0.0687 | | cosine_recall@1 | 0.2952 | | cosine_recall@3 | 0.5241 | | cosine_recall@5 | 0.5843 | | cosine_recall@10 | 0.6867 | | cosine_ndcg@10 | 0.4908 | | cosine_mrr@10 | 0.4284 | | **cosine_map@100** | **0.4358** | #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.259 | | cosine_accuracy@3 | 0.506 | | cosine_accuracy@5 | 0.5783 | | cosine_accuracy@10 | 0.6446 | | cosine_precision@1 | 0.259 | | cosine_precision@3 | 0.1687 | | cosine_precision@5 | 0.1157 | | cosine_precision@10 | 0.0645 | | cosine_recall@1 | 0.259 | | cosine_recall@3 | 0.506 | | cosine_recall@5 | 0.5783 | | cosine_recall@10 | 0.6446 | | cosine_ndcg@10 | 0.4548 | | cosine_mrr@10 | 0.3935 | | **cosine_map@100** | **0.4034** | #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.2711 | | cosine_accuracy@3 | 0.4699 | | cosine_accuracy@5 | 0.5663 | | cosine_accuracy@10 | 0.6145 | | cosine_precision@1 | 0.2711 | | cosine_precision@3 | 0.1566 | | cosine_precision@5 | 0.1133 | | cosine_precision@10 | 0.0614 | | cosine_recall@1 | 0.2711 | | cosine_recall@3 | 0.4699 | | cosine_recall@5 | 0.5663 | | cosine_recall@10 | 0.6145 | | cosine_ndcg@10 | 0.4443 | | cosine_mrr@10 | 0.3894 | | **cosine_map@100** | **0.3989** | #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.2169 | | cosine_accuracy@3 | 0.4217 | | cosine_accuracy@5 | 0.5181 | | cosine_accuracy@10 | 0.5843 | | cosine_precision@1 | 0.2169 | | cosine_precision@3 | 0.1406 | | cosine_precision@5 | 0.1036 | | cosine_precision@10 | 0.0584 | | cosine_recall@1 | 0.2169 | | cosine_recall@3 | 0.4217 | | cosine_recall@5 | 0.5181 | | cosine_recall@10 | 0.5843 | | cosine_ndcg@10 | 0.3964 | | cosine_mrr@10 | 0.3365 | | **cosine_map@100** | **0.3466** |   ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 1,490 training samples * Columns: <code>positive</code>, <code>anchor</code>, and <code>negative</code> * Approximate statistics based on the first 1000 samples: | | positive | anchor | negative | |:--------|:----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------| | type | string | string | string | | details | <ul><li>min: 9 tokens</li><li>mean: 21.02 tokens</li><li>max: 64 tokens</li></ul> | <ul><li>min: 23 tokens</li><li>mean: 375.16 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 10 tokens</li><li>mean: 17.51 tokens</li><li>max: 31 tokens</li></ul> | * Samples: | positive | anchor | negative | |:-----------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------| | <code>What details can you provide about the mlflow_training_pipeline runs listed in the ZenML documentation?</code> | <code>mlflow_training_pipeline', ┃┃ │ │ │ 'zenml_pipeline_run_uuid': 'a5d4faae-ef70-48f2-9893-6e65d5e51e98', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.005'} ┃ ┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨ ┃ tensorflow-mnist-model │ 2 │ Run #2 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_09_08_467212', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃ ┃ │ │ │ 'zenml_pipeline_run_uuid': '11858dcf-3e47-4b1a-82c5-6fa25ba4e037', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.003'} ┃ ┠────────────────────────┼───────────────┼─────────────────────────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┨ ┃ tensorflow-mnist-model │ 1 │ Run #1 of the mlflow_training_pipeline. │ {'zenml_version': '0.34.0', 'zenml_run_name': 'mlflow_training_pipeline-2023_03_01-08_08_52_398499', 'zenml_pipeline_name': 'mlflow_training_pipeline', ┃ ┃ │ │ │ 'zenml_pipeline_run_uuid': '29fb22c1-6e0b-4431-9e04-226226506d16', 'zenml_workspace': '10e060b3-2f7e-463d-9ec8-3a211ef4e1f6', 'epochs': '5', 'optimizer': 'Adam', 'lr': '0.001'} ┃</code> | <code>Can you explain how to configure the TensorFlow settings for a different project?</code> | | <code>How do you register a GCP Service Connector that uses account impersonation to access the zenml-bucket-sl GCS bucket?</code> | <code>esource-id zenml-bucket-sl Example Command OutputError: Service connector 'gcp-empty-sa' verification failed: connector authorization failure: failed to fetch GCS bucket zenml-bucket-sl: 403 GET https://storage.googleapis.com/storage/v1/b/zenml-bucket-sl?projection=noAcl&prettyPrint=false: [email protected] does not have storage.buckets.get access to the Google Cloud Storage bucket. Permission 'storage.buckets.get' denied on resource (or it may not exist). Next, we'll register a GCP Service Connector that actually uses account impersonation to access the zenml-bucket-sl GCS bucket and verify that it can actually access the bucket: zenml service-connector register gcp-impersonate-sa --type gcp --auth-method impersonation --service_account_json=@[email protected] --project_id=zenml-core --target_principal=zenml-bucket-sl@zenml-core.iam.gserviceaccount.com --resource-type gcs-bucket --resource-id gs://zenml-bucket-sl Example Command Output Expanding argument value service_account_json to contents of file /home/stefan/aspyre/src/zenml/[email protected]. Successfully registered service connector `gcp-impersonate-sa` with access to the following resources: ┏━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━┓ ┃ RESOURCE TYPE │ RESOURCE NAMES ┃ ┠───────────────┼──────────────────────┨ ┃ 📦 gcs-bucket │ gs://zenml-bucket-sl ┃ ┗━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━┛ External Account (GCP Workload Identity) Use GCP workload identity federation to authenticate to GCP services using AWS IAM credentials, Azure Active Directory credentials or generic OIDC tokens.</code> | <code>What is the process for setting up a ZenML pipeline using AWS IAM credentials?</code> | | <code>Can you explain how data validation helps in detecting data drift and model drift in ZenML pipelines?</code> | <code>of your models at different stages of development.if you have pipelines that regularly ingest new data, you should use data validation to run regular data integrity checks to signal problems before they are propagated downstream. in continuous training pipelines, you should use data validation techniques to compare new training data against a data reference and to compare the performance of newly trained models against previous ones. when you have pipelines that automate batch inference or if you regularly collect data used as input in online inference, you should use data validation to run data drift analyses and detect training-serving skew, data drift and model drift. Data Validator Flavors Data Validator are optional stack components provided by integrations. The following table lists the currently available Data Validators and summarizes their features and the data types and model types that they can be used with in ZenML pipelines: Data Validator Validation Features Data Types Model Types Notes Flavor/Integration Deepchecks data quality data drift model drift model performance tabular: pandas.DataFrame CV: torch.utils.data.dataloader.DataLoader tabular: sklearn.base.ClassifierMixin CV: torch.nn.Module Add Deepchecks data and model validation tests to your pipelines deepchecks Evidently data quality data drift model drift model performance tabular: pandas.DataFrame N/A Use Evidently to generate a variety of data quality and data/model drift reports and visualizations evidently Great Expectations data profiling data quality tabular: pandas.DataFrame N/A Perform data testing, documentation and profiling with Great Expectations great_expectations Whylogs/WhyLabs data drift tabular: pandas.DataFrame N/A Generate data profiles with whylogs and upload them to WhyLabs whylogs If you would like to see the available flavors of Data Validator, you can use the command: zenml data-validator flavor list How to use it</code> | <code>What are the best practices for deploying web applications using Docker and Kubernetes?</code> | * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "TripletLoss", "matryoshka_dims": [ 384, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `gradient_accumulation_steps`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 4 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 16 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 4 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: True - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs | Epoch | Step | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_384_cosine_map@100 | dim_64_cosine_map@100 | |:-------:|:-----:|:----------------------:|:----------------------:|:----------------------:|:---------------------:| | 0.6667 | 1 | 0.3884 | 0.4332 | 0.4464 | 0.3140 | | **2.0** | **3** | **0.4064** | **0.4195** | **0.4431** | **0.3553** | | 2.6667 | 4 | 0.3989 | 0.4034 | 0.4358 | 0.3466 | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.3.1+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### TripletLoss ```bibtex @misc{hermans2017defense, title={In Defense of the Triplet Loss for Person Re-Identification}, author={Alexander Hermans and Lucas Beyer and Bastian Leibe}, year={2017}, eprint={1703.07737}, archivePrefix={arXiv}, primaryClass={cs.CV} } ```

{"base_model": "Snowflake/snowflake-arctic-embed-m", "datasets": [], "language": ["en"], "library_name": "sentence-transformers", "license": "apache-2.0", "metrics": ["cosine_accuracy@1", "cosine_accuracy@3", "cosine_accuracy@5", "cosine_accuracy@10", "cosine_precision@1", "cosine_precision@3", "cosine_precision@5", "cosine_precision@10", "cosine_recall@1", "cosine_recall@3", "cosine_recall@5", "cosine_recall@10", "cosine_ndcg@10", "cosine_mrr@10", "cosine_map@100"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:1490", "loss:MatryoshkaLoss", "loss:TripletLoss"], "widget": [{"source_sentence": "Where is the global configuration directory located in ZenML's default setup?", "sentences": ["'default' ...\n\nCreating default user 'default' ...Creating default stack for user 'default' in workspace default...\n\nActive workspace not set. Setting it to the default.\n\nThe active stack is not set. Setting the active stack to the default workspace stack.\n\nUsing the default store for the global config.\n\nUnable to find ZenML repository in your current working directory (/tmp/folder) or any parent directories. If you want to use an existing repository which is in a different location, set the environment variable 'ZENML_REPOSITORY_PATH'. If you want to create a new repository, run zenml init.\n\nRunning without an active repository root.\n\nUsing the default local database.\n\nRunning with active workspace: 'default' (global)\n\n┏━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━┓\n\n┃ ACTIVE │ STACK NAME │ SHARED │ OWNER │ ARTIFACT_STORE │ ORCHESTRATOR ┃\n\n┠────────┼────────────┼────────┼─────────┼────────────────┼──────────────┨\n\n┃ 👉 │ default │ ❌ │ default │ default │ default ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━┛\n\nThe following is an example of the layout of the global config directory immediately after initialization:\n\n/home/stefan/.config/zenml <- Global Config Directory\n\n├── config.yaml <- Global Configuration Settings\n\n└── local_stores <- Every Stack component that stores information\n\n| locally will have its own subdirectory here.\n\n├── a1a0d3d0-d552-4a80-be09-67e5e29be8ee <- e.g. Local Store path for the\n\n| `default` local Artifact Store\n\n└── default_zen_store\n\n└── zenml.db <- SQLite database where ZenML data (stacks,\n\ncomponents, etc) are stored by default.\n\nAs shown above, the global config directory stores the following information:", "How do you configure the network settings on a Linux server?", "Reranking for better retrieval\n\nAdd reranking to your RAG inference for better retrieval performance.\n\nRerankers are a crucial component of retrieval systems that use LLMs. They help improve the quality of the retrieved documents by reordering them based on additional features or scores. In this section, we'll explore how to add a reranker to your RAG inference pipeline in ZenML.\n\nIn previous sections, we set up the overall workflow, from data ingestion and preprocessing to embeddings generation and retrieval. We then set up some basic evaluation metrics to assess the performance of our retrieval system. A reranker is a way to squeeze a bit of extra performance out of the system by reordering the retrieved documents based on additional features or scores.\n\nAs you can see, reranking is an optional addition we make to what we've already set up. It's not strictly necessary, but it can help improve the relevance and quality of the retrieved documents, which in turn can lead to better responses from the LLM. Let's dive in!\n\nPreviousEvaluation in practice\n\nNextUnderstanding reranking\n\nLast updated 1 month ago"]}, {"source_sentence": "Where can I find the instructions to enable CUDA for GPU-backed hardware in ZenML SDK Docs?", "sentences": ["Migration guide 0.39.1 → 0.41.0\n\nHow to migrate your ZenML pipelines and steps from version <=0.39.1 to 0.41.0.\n\nZenML versions 0.40.0 to 0.41.0 introduced a new and more flexible syntax to define ZenML steps and pipelines. This page contains code samples that show you how to upgrade your steps and pipelines to the new syntax.\n\nNewer versions of ZenML still work with pipelines and steps defined using the old syntax, but the old syntax is deprecated and will be removed in the future.\n\nOverview\n\nfrom typing import Optional\n\nfrom zenml.steps import BaseParameters, Output, StepContext, step\n\nfrom zenml.pipelines import pipeline\n\n# Define a Step\n\nclass MyStepParameters(BaseParameters):\n\nparam_1: int\n\nparam_2: Optional[float] = None\n\n@step\n\ndef my_step(\n\nparams: MyStepParameters, context: StepContext,\n\n) -> Output(int_output=int, str_output=str):\n\nresult = int(params.param_1 * (params.param_2 or 1))\n\nresult_uri = context.get_output_artifact_uri()\n\nreturn result, result_uri\n\n# Run the Step separately\n\nmy_step.entrypoint()\n\n# Define a Pipeline\n\n@pipeline\n\ndef my_pipeline(my_step):\n\nmy_step()\n\nstep_instance = my_step(params=MyStepParameters(param_1=17))\n\npipeline_instance = my_pipeline(my_step=step_instance)\n\n# Configure and run the Pipeline\n\npipeline_instance.configure(enable_cache=False)\n\nschedule = Schedule(...)\n\npipeline_instance.run(schedule=schedule)\n\n# Fetch the Pipeline Run\n\nlast_run = pipeline_instance.get_runs()[0]\n\nint_output = last_run.get_step[\"my_step\"].outputs[\"int_output\"].read()\n\nfrom typing import Annotated, Optional, Tuple\n\nfrom zenml import get_step_context, pipeline, step\n\nfrom zenml.client import Client\n\n# Define a Step\n\n@step\n\ndef my_step(\n\nparam_1: int, param_2: Optional[float] = None\n\n) -> Tuple[Annotated[int, \"int_output\"], Annotated[str, \"str_output\"]]:\n\nresult = int(param_1 * (param_2 or 1))\n\nresult_uri = get_step_context().get_output_artifact_uri()\n\nreturn result, result_uri\n\n# Run the Step separately\n\nmy_step()\n\n# Define a Pipeline\n\n@pipeline", "How do I integrate Google Cloud VertexAI into my existing Kubernetes cluster?", " SDK Docs .\n\nEnabling CUDA for GPU-backed hardwareNote that if you wish to use this step operator to run steps on a GPU, you will need to follow the instructions on this page to ensure that it works. It requires adding some extra settings customization and is essential to enable CUDA for the GPU to give its full acceleration.\n\nPreviousStep Operators\n\nNextGoogle Cloud VertexAI\n\nLast updated 19 days ago"]}, {"source_sentence": "What are the special metadata types supported by ZenML and how are they used?", "sentences": ["Special Metadata Types\n\nTracking your metadata.\n\nZenML supports several special metadata types to capture specific kinds of information. Here are examples of how to use the special types Uri, Path, DType, and StorageSize:\n\nfrom zenml.metadata.metadata_types import StorageSize, DType\n\nfrom zenml import log_artifact_metadata\n\nlog_artifact_metadata(\n\nmetadata={\n\n\"dataset_source\": Uri(\"gs://my-bucket/datasets/source.csv\"),\n\n\"preprocessing_script\": Path(\"/scripts/preprocess.py\"),\n\n\"column_types\": {\n\n\"age\": DType(\"int\"),\n\n\"income\": DType(\"float\"),\n\n\"score\": DType(\"int\")\n\n},\n\n\"processed_data_size\": StorageSize(2500000)\n\nIn this example:\n\nUri is used to indicate a dataset source URI.\n\nPath is used to specify the filesystem path to a preprocessing script.\n\nDType is used to describe the data types of specific columns.\n\nStorageSize is used to indicate the size of the processed data in bytes.\n\nThese special types help standardize the format of metadata and ensure that it is logged in a consistent and interpretable manner.\n\nPreviousGroup metadata\n\nNextFetch metadata within steps\n\nLast updated 19 days ago", "Configure a code repository\n\nConnect a Git repository to ZenML to track code changes and collaborate on MLOps projects.\n\nThroughout the lifecycle of a MLOps pipeline, it can get quite tiresome to always wait for a Docker build every time after running a pipeline (even if the local Docker cache is used). However, there is a way to just have one pipeline build and keep reusing it until a change to the pipeline environment is made: by connecting a code repository.\n\nWith ZenML, connecting to a Git repository optimizes the Docker build processes. It also has the added bonus of being a better way of managing repository changes and enabling better code collaboration. Here is how the flow changes when running a pipeline:\n\nYou trigger a pipeline run on your local machine. ZenML parses the @pipeline function to determine the necessary steps.\n\nThe local client requests stack information from the ZenML server, which responds with the cloud stack configuration.\n\nThe local client detects that we're using a code repository and requests the information from the git repo.\n\nInstead of building a new Docker image, the client checks if an existing image can be reused based on the current Git commit hash and other environment metadata.\n\nThe client initiates a run in the orchestrator, which sets up the execution environment in the cloud, such as a VM.\n\nThe orchestrator downloads the code directly from the Git repository and uses the existing Docker image to run the pipeline steps.\n\nPipeline steps execute, storing artifacts in the cloud-based artifact store.\n\nThroughout the execution, the pipeline run status and metadata are reported back to the ZenML server.\n\nBy connecting a Git repository, you avoid redundant builds and make your MLOps processes more efficient. Your team can work on the codebase simultaneously, with ZenML handling the version tracking and ensuring that the correct code version is always used for each run.\n\nCreating a GitHub Repository", "Can you explain the process of setting up a virtual environment in Python?"]}, {"source_sentence": "What are the benefits of deploying stack components directly from the ZenML CLI?", "sentences": ["─────────────────────────────────────────────────┨┃ RESOURCE TYPES │ 🔵 gcp-generic, 📦 gcs-bucket, 🌀 kubernetes-cluster, 🐳 docker-registry ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ RESOURCE NAME │ <multiple> ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ SECRET ID │ 4694de65-997b-4929-8831-b49d5e067b97 ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ SESSION DURATION │ N/A ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ EXPIRES IN │ 59m46s ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ OWNER │ default ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ WORKSPACE │ default ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ SHARED │ ➖ ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ CREATED_AT │ 2023-05-19 09:04:33.557126 ┃\n\n┠──────────────────┼──────────────────────────────────────────────────────────────────────────┨\n\n┃ UPDATED_AT │ 2023-05-19 09:04:33.557127 ┃\n\n┗━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓", "How do you set up a custom service account for Vertex AI?", "⚒️Manage stacks\n\nDeploying your stack components directly from the ZenML CLI\n\nThe first step in running your pipelines on remote infrastructure is to deploy all the components that you would need, like an MLflow tracking server, a Seldon Core model deployer, and more to your cloud.\n\nThis can bring plenty of benefits like scalability, reliability, and collaboration. ZenML eases the path to production by providing a seamless way for all tools to interact with others through the use of abstractions. However, one of the most painful parts of this process, from what we see on our Slack and in general, is the deployment of these stack components.\n\nDeploying and managing MLOps tools is tricky 😭😵‍💫\n\nIt is not trivial to set up all the different tools that you might need for your pipeline.\n\n🌈 Each tool comes with a certain set of requirements. For example, a Kubeflow installation will require you to have a Kubernetes cluster, and so would a Seldon Core deployment.\n\n🤔 Figuring out the defaults for infra parameters is not easy. Even if you have identified the backing infra that you need for a stack component, setting up reasonable defaults for parameters like instance size, CPU, memory, etc., needs a lot of experimentation to figure out.\n\n🚧 Many times, standard tool installations don't work out of the box. For example, to run a custom pipeline in Vertex AI, it is not enough to just run an imported pipeline. You might also need a custom service account that is configured to perform tasks like reading secrets from your secret store or talking to other GCP services that your pipeline might need.\n\n🔐 Some tools need an additional layer of installations to enable a more secure, production-grade setup. For example, a standard MLflow tracking server deployment comes without an authentication frontend which might expose all of your tracking data to the world if deployed as-is."]}, {"source_sentence": "What is the expiration time for the GCP OAuth2 token in the ZenML configuration?", "sentences": ["━━━━━┛\n\nConfiguration\n\n┏━━━━━━━━━━━━┯━━━━━━━━━━━━┓┃ PROPERTY │ VALUE ┃\n\n┠────────────┼────────────┨\n\n┃ project_id │ zenml-core ┃\n\n┠────────────┼────────────┨\n\n┃ token │ [HIDDEN] ┃\n\n┗━━━━━━━━━━━━┷━━━━━━━━━━━━┛\n\nNote the temporary nature of the Service Connector. It will expire and become unusable in 1 hour:\n\nzenml service-connector list --name gcp-oauth2-token\n\nExample Command Output\n\n┏━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━┓\n\n┃ ACTIVE │ NAME │ ID │ TYPE │ RESOURCE TYPES │ RESOURCE NAME │ SHARED │ OWNER │ EXPIRES IN │ LABELS ┃\n\n┠────────┼──────────────────┼──────────────────────────────────────┼────────┼───────────────────────┼───────────────┼────────┼─────────┼────────────┼────────┨\n\n┃ │ gcp-oauth2-token │ ec4d7d85-c71c-476b-aa76-95bf772c90da │ 🔵 gcp │ 🔵 gcp-generic │ <multiple> │ ➖ │ default │ 59m35s │ ┃\n\n┃ │ │ │ │ 📦 gcs-bucket │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🌀 kubernetes-cluster │ │ │ │ │ ┃\n\n┃ │ │ │ │ 🐳 docker-registry │ │ │ │ │ ┃\n\n┗━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━┛\n\nAuto-configuration\n\nThe GCP Service Connector allows auto-discovering and fetching credentials and configuration set up by the GCP CLI on your local host.", "Hugging Face\n\nDeploying models to Huggingface Inference Endpoints with Hugging Face :hugging_face:.\n\nHugging Face Inference Endpoints provides a secure production solution to easily deploy any transformers, sentence-transformers, and diffusers models on a dedicated and autoscaling infrastructure managed by Hugging Face. An Inference Endpoint is built from a model from the Hub.\n\nThis service provides dedicated and autoscaling infrastructure managed by Hugging Face, allowing you to deploy models without dealing with containers and GPUs.\n\nWhen to use it?\n\nYou should use Hugging Face Model Deployer:\n\nif you want to deploy Transformers, Sentence-Transformers, or Diffusion models on dedicated and secure infrastructure.\n\nif you prefer a fully-managed production solution for inference without the need to handle containers and GPUs.\n\nif your goal is to turn your models into production-ready APIs with minimal infrastructure or MLOps involvement\n\nCost-effectiveness is crucial, and you want to pay only for the raw compute resources you use.\n\nEnterprise security is a priority, and you need to deploy models into secure offline endpoints accessible only via a direct connection to your Virtual Private Cloud (VPCs).\n\nIf you are looking for a more easy way to deploy your models locally, you can use the MLflow Model Deployer flavor.\n\nHow to deploy it?\n\nThe Hugging Face Model Deployer flavor is provided by the Hugging Face ZenML integration, so you need to install it on your local machine to be able to deploy your models. You can do this by running the following command:\n\nzenml integration install huggingface -y\n\nTo register the Hugging Face model deployer with ZenML you need to run the following command:\n\nzenml model-deployer register <MODEL_DEPLOYER_NAME> --flavor=huggingface --token=<YOUR_HF_TOKEN> --namespace=<YOUR_HF_NAMESPACE>\n\nHere,\n\ntoken parameter is the Hugging Face authentication token. It can be managed through Hugging Face settings.", "Can you list the steps to set up a Docker registry on a Kubernetes cluster?"]}], "model-index": [{"name": "zenml/finetuned-snowflake-arctic-embed-m", "results": [{"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 384", "type": "dim_384"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.29518072289156627, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.5240963855421686, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5843373493975904, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.6867469879518072, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.29518072289156627, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.17469879518072293, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.11686746987951804, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.0686746987951807, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.29518072289156627, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.5240963855421686, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5843373493975904, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.6867469879518072, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.4908042072911187, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.42844234079173843, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.43576329240226386, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 256", "type": "dim_256"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.25903614457831325, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.5060240963855421, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5783132530120482, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.6445783132530121, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.25903614457831325, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.1686746987951807, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.11566265060240961, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.0644578313253012, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.25903614457831325, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.5060240963855421, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5783132530120482, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.6445783132530121, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.4548319777111225, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.39346194301013593, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.40343211538391555, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 128", "type": "dim_128"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.2710843373493976, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.46987951807228917, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5662650602409639, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.6144578313253012, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.2710843373493976, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.1566265060240964, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.11325301204819276, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.061445783132530116, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.2710843373493976, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.46987951807228917, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5662650602409639, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.6144578313253012, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.44433019669319024, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.3893574297188756, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.3989315479842741, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 64", "type": "dim_64"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.21686746987951808, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.42168674698795183, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.5180722891566265, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.5843373493975904, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.21686746987951808, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.14056224899598396, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.10361445783132528, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.05843373493975902, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.21686746987951808, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.42168674698795183, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.5180722891566265, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.5843373493975904, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.39639025659520544, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.3364529546758464, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.34658882510541217, "name": "Cosine Map@100"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,574

debarghabhattofficial/t5-small-squad-qg-a2c-spt-valid

debarghabhattofficial

text2text-generation

[ "transformers", "pytorch", "tensorboard", "t5", "text2text-generation", "generated_from_trainer", "dataset:qg_squad", "license:cc-by-4.0", "model-index", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2023-04-03T11:39:31Z

2023-04-03T19:25:27+00:00

12

0

--- datasets: - qg_squad license: cc-by-4.0 metrics: - bleu tags: - generated_from_trainer model-index: - name: t5-small-squad-qg-a2c-spt-valid results: - task: type: text2text-generation name: Sequence-to-sequence Language Modeling dataset: name: qg_squad type: qg_squad config: qg_squad split: test args: qg_squad metrics: - type: bleu value: 0.1856298695745541 name: Bleu ---  # t5-small-squad-qg-a2c-spt-valid This model is a fine-tuned version of [lmqg/t5-small-squad-qg](https://huggingface.co/lmqg/t5-small-squad-qg) on the qg_squad dataset. It achieves the following results on the evaluation set: - Loss: 3.5585 - Bleu: 0.1856 - Precisions: [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] - Brevity Penalty: 0.9231 - Length Ratio: 0.9259 - Translation Length: 126899 - Reference Length: 137056 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 - label_smoothing_factor: 0.15 ### Training results | Training Loss | Epoch | Step | Validation Loss | Bleu | Precisions | Brevity Penalty | Length Ratio | Translation Length | Reference Length | |:-------------:|:-----:|:-----:|:---------------:|:------:|:-----------------------------------------------------------------------------------:|:---------------:|:------------:|:------------------:|:----------------:| | 3.4717 | 1.0 | 1184 | 3.5703 | 0.1850 | [0.4884210026960997, 0.23740423378300554, 0.14702360671696277, 0.09591845720324058] | 0.9198 | 0.9228 | 126479 | 137056 | | 3.4432 | 2.0 | 2368 | 3.5676 | 0.1847 | [0.4899809765377299, 0.23739313808702955, 0.14709099076226004, 0.09610180163262601] | 0.9173 | 0.9205 | 126160 | 137056 | | 3.4207 | 3.0 | 3552 | 3.5654 | 0.1855 | [0.48690609948692964, 0.236654650074526, 0.14669770766719153, 0.09533838196460138] | 0.9260 | 0.9286 | 127273 | 137056 | | 3.4017 | 4.0 | 4736 | 3.5575 | 0.1861 | [0.4907433036243861, 0.23905491743183327, 0.14802083840498564, 0.09654473782730295] | 0.9195 | 0.9226 | 126449 | 137056 | | 3.3862 | 5.0 | 5920 | 3.5540 | 0.1851 | [0.4916027385306181, 0.23877172085201795, 0.14769450336757936, 0.09608281170511601] | 0.9164 | 0.9197 | 126053 | 137056 | | 3.3715 | 6.0 | 7104 | 3.5619 | 0.1847 | [0.4897172642552519, 0.23742624822429256, 0.14650127350144848, 0.09495653320731078] | 0.9209 | 0.9239 | 126620 | 137056 | | 3.3602 | 7.0 | 8288 | 3.5581 | 0.1857 | [0.49199648336329865, 0.2390627732121, 0.14782006380301063, 0.09637410897534923] | 0.9180 | 0.9212 | 126257 | 137056 | | 3.3523 | 8.0 | 9472 | 3.5575 | 0.1856 | [0.4896288812767368, 0.23802266135985578, 0.14728396021137705, 0.09588544697859817] | 0.9215 | 0.9244 | 126698 | 137056 | | 3.3439 | 9.0 | 10656 | 3.5582 | 0.1862 | [0.4919672196048933, 0.23971752696254087, 0.14848694668474074, 0.09658739962940087] | 0.9183 | 0.9215 | 126295 | 137056 | | 3.3395 | 10.0 | 11840 | 3.5585 | 0.1856 | [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] | 0.9231 | 0.9259 | 126899 | 137056 | ### Framework versions - Transformers 4.27.4 - Pytorch 1.9.0 - Datasets 2.9.0 - Tokenizers 0.13.2

null

Non_BioNLP

# t5-small-squad-qg-a2c-spt-valid This model is a fine-tuned version of [lmqg/t5-small-squad-qg](https://huggingface.co/lmqg/t5-small-squad-qg) on the qg_squad dataset. It achieves the following results on the evaluation set: - Loss: 3.5585 - Bleu: 0.1856 - Precisions: [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] - Brevity Penalty: 0.9231 - Length Ratio: 0.9259 - Translation Length: 126899 - Reference Length: 137056 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 - label_smoothing_factor: 0.15 ### Training results | Training Loss | Epoch | Step | Validation Loss | Bleu | Precisions | Brevity Penalty | Length Ratio | Translation Length | Reference Length | |:-------------:|:-----:|:-----:|:---------------:|:------:|:-----------------------------------------------------------------------------------:|:---------------:|:------------:|:------------------:|:----------------:| | 3.4717 | 1.0 | 1184 | 3.5703 | 0.1850 | [0.4884210026960997, 0.23740423378300554, 0.14702360671696277, 0.09591845720324058] | 0.9198 | 0.9228 | 126479 | 137056 | | 3.4432 | 2.0 | 2368 | 3.5676 | 0.1847 | [0.4899809765377299, 0.23739313808702955, 0.14709099076226004, 0.09610180163262601] | 0.9173 | 0.9205 | 126160 | 137056 | | 3.4207 | 3.0 | 3552 | 3.5654 | 0.1855 | [0.48690609948692964, 0.236654650074526, 0.14669770766719153, 0.09533838196460138] | 0.9260 | 0.9286 | 127273 | 137056 | | 3.4017 | 4.0 | 4736 | 3.5575 | 0.1861 | [0.4907433036243861, 0.23905491743183327, 0.14802083840498564, 0.09654473782730295] | 0.9195 | 0.9226 | 126449 | 137056 | | 3.3862 | 5.0 | 5920 | 3.5540 | 0.1851 | [0.4916027385306181, 0.23877172085201795, 0.14769450336757936, 0.09608281170511601] | 0.9164 | 0.9197 | 126053 | 137056 | | 3.3715 | 6.0 | 7104 | 3.5619 | 0.1847 | [0.4897172642552519, 0.23742624822429256, 0.14650127350144848, 0.09495653320731078] | 0.9209 | 0.9239 | 126620 | 137056 | | 3.3602 | 7.0 | 8288 | 3.5581 | 0.1857 | [0.49199648336329865, 0.2390627732121, 0.14782006380301063, 0.09637410897534923] | 0.9180 | 0.9212 | 126257 | 137056 | | 3.3523 | 8.0 | 9472 | 3.5575 | 0.1856 | [0.4896288812767368, 0.23802266135985578, 0.14728396021137705, 0.09588544697859817] | 0.9215 | 0.9244 | 126698 | 137056 | | 3.3439 | 9.0 | 10656 | 3.5582 | 0.1862 | [0.4919672196048933, 0.23971752696254087, 0.14848694668474074, 0.09658739962940087] | 0.9183 | 0.9215 | 126295 | 137056 | | 3.3395 | 10.0 | 11840 | 3.5585 | 0.1856 | [0.4899881007730557, 0.23798056024064962, 0.14699694604682728, 0.09541131612394267] | 0.9231 | 0.9259 | 126899 | 137056 | ### Framework versions - Transformers 4.27.4 - Pytorch 1.9.0 - Datasets 2.9.0 - Tokenizers 0.13.2

{"datasets": ["qg_squad"], "license": "cc-by-4.0", "metrics": ["bleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "t5-small-squad-qg-a2c-spt-valid", "results": [{"task": {"type": "text2text-generation", "name": "Sequence-to-sequence Language Modeling"}, "dataset": {"name": "qg_squad", "type": "qg_squad", "config": "qg_squad", "split": "test", "args": "qg_squad"}, "metrics": [{"type": "bleu", "value": 0.1856298695745541, "name": "Bleu"}]}]}]}

task

[ "TRANSLATION" ]

44,575

Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF

Triangle104

text-generation

[ "transformers", "gguf", "safetensors", "onnx", "transformers.js", "llama-cpp", "gguf-my-repo", "text-generation", "en", "base_model:HuggingFaceTB/SmolLM2-135M-Instruct", "base_model:quantized:HuggingFaceTB/SmolLM2-135M-Instruct", "license:apache-2.0", "endpoints_compatible", "region:us", "conversational" ]

2024-11-07T05:02:37Z

2025-01-14T07:12:49+00:00

7

0

--- base_model: HuggingFaceTB/SmolLM2-135M-Instruct language: - en library_name: transformers license: apache-2.0 pipeline_tag: text-generation tags: - safetensors - onnx - transformers.js - llama-cpp - gguf-my-repo --- # Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF This model was converted to GGUF format from [`HuggingFaceTB/SmolLM2-135M-Instruct`](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) for more details on the model. --- Model details: - SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters. They are capable of solving a wide range of tasks while being lightweight enough to run on-device. SmolLM2 demonstrates significant advances over its predecessor SmolLM1, particularly in instruction following, knowledge, reasoning. The 135M model was trained on 2 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new filtered datasets we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback. The instruct model additionally supports tasks such as text rewriting, summarization and function calling (for the 1.7B) thanks to datasets developed by Argilla such as Synth-APIGen-v0.1. You can find the SFT dataset here: https://huggingface.co/datasets/HuggingFaceTB/smol-smoltalk and finetuning code at https://github.com/huggingface/alignment-handbook/tree/main/recipes/smollm2 How to use Transformers pip install transformers from transformers import AutoModelForCausalLM, AutoTokenizer checkpoint = "HuggingFaceTB/SmolLM2-135M-Instruct" device = "cuda" # for GPU usage or "cpu" for CPU usage tokenizer = AutoTokenizer.from_pretrained(checkpoint) # for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")` model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device) messages = [{"role": "user", "content": "What is gravity?"}] input_text=tokenizer.apply_chat_template(messages, tokenize=False) print(input_text) inputs = tokenizer.encode(input_text, return_tensors="pt").to(device) outputs = model.generate(inputs, max_new_tokens=50, temperature=0.2, top_p=0.9, do_sample=True) print(tokenizer.decode(outputs[0])) Chat in TRL You can also use the TRL CLI to chat with the model from the terminal: pip install trl trl chat --model_name_or_path HuggingFaceTB/SmolLM2-135M-Instruct --device cpu --- ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ```

null

Non_BioNLP

# Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF This model was converted to GGUF format from [`HuggingFaceTB/SmolLM2-135M-Instruct`](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct) for more details on the model. --- Model details: - SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters. They are capable of solving a wide range of tasks while being lightweight enough to run on-device. SmolLM2 demonstrates significant advances over its predecessor SmolLM1, particularly in instruction following, knowledge, reasoning. The 135M model was trained on 2 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new filtered datasets we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback. The instruct model additionally supports tasks such as text rewriting, summarization and function calling (for the 1.7B) thanks to datasets developed by Argilla such as Synth-APIGen-v0.1. You can find the SFT dataset here: https://huggingface.co/datasets/HuggingFaceTB/smol-smoltalk and finetuning code at https://github.com/huggingface/alignment-handbook/tree/main/recipes/smollm2 How to use Transformers pip install transformers from transformers import AutoModelForCausalLM, AutoTokenizer checkpoint = "HuggingFaceTB/SmolLM2-135M-Instruct" device = "cuda" # for GPU usage or "cpu" for CPU usage tokenizer = AutoTokenizer.from_pretrained(checkpoint) # for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")` model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device) messages = [{"role": "user", "content": "What is gravity?"}] input_text=tokenizer.apply_chat_template(messages, tokenize=False) print(input_text) inputs = tokenizer.encode(input_text, return_tensors="pt").to(device) outputs = model.generate(inputs, max_new_tokens=50, temperature=0.2, top_p=0.9, do_sample=True) print(tokenizer.decode(outputs[0])) Chat in TRL You can also use the TRL CLI to chat with the model from the terminal: pip install trl trl chat --model_name_or_path HuggingFaceTB/SmolLM2-135M-Instruct --device cpu --- ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo Triangle104/SmolLM2-135M-Instruct-Q8_0-GGUF --hf-file smollm2-135m-instruct-q8_0.gguf -c 2048 ```

{"base_model": "HuggingFaceTB/SmolLM2-135M-Instruct", "language": ["en"], "library_name": "transformers", "license": "apache-2.0", "pipeline_tag": "text-generation", "tags": ["safetensors", "onnx", "transformers.js", "llama-cpp", "gguf-my-repo"]}

task

[ "SUMMARIZATION" ]

44,576

fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391

fine-tuned

feature-extraction

[ "sentence-transformers", "safetensors", "bert", "feature-extraction", "sentence-similarity", "mteb", "custom_code", "en", "dataset:fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391", "dataset:allenai/c4", "license:apache-2.0", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

2024-05-28T23:43:36Z

2024-05-28T23:43:55+00:00

7

0

--- datasets: - fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391 - allenai/c4 language: - en - en license: apache-2.0 pipeline_tag: feature-extraction tags: - sentence-transformers - feature-extraction - sentence-similarity - mteb --- This model is a fine-tuned version of [**jinaai/jina-embeddings-v2-base-en**](https://huggingface.co/jinaai/jina-embeddings-v2-base-en) designed for the following use case: None ## How to Use This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started: ```python from sentence_transformers import SentenceTransformer from sentence_transformers.util import cos_sim model = SentenceTransformer( 'fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391', trust_remote_code=True ) embeddings = model.encode([ 'first text to embed', 'second text to embed' ]) print(cos_sim(embeddings[0], embeddings[1])) ```

null

Non_BioNLP

This model is a fine-tuned version of [**jinaai/jina-embeddings-v2-base-en**](https://huggingface.co/jinaai/jina-embeddings-v2-base-en) designed for the following use case: None ## How to Use This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started: ```python from sentence_transformers import SentenceTransformer from sentence_transformers.util import cos_sim model = SentenceTransformer( 'fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391', trust_remote_code=True ) embeddings = model.encode([ 'first text to embed', 'second text to embed' ]) print(cos_sim(embeddings[0], embeddings[1])) ```

{"datasets": ["fine-tuned/FiQA2018-512-192-gpt-4o-2024-05-13-87401391", "allenai/c4"], "language": ["en", "en"], "license": "apache-2.0", "pipeline_tag": "feature-extraction", "tags": ["sentence-transformers", "feature-extraction", "sentence-similarity", "mteb"]}

task

[ "TEXT_CLASSIFICATION" ]

44,577

zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF

zzz-z-z-z

translation

[ "gguf", "llama-cpp", "gguf-my-repo", "translation", "en", "de", "fr", "zh", "pt", "nl", "ru", "ko", "it", "es", "base_model:Unbabel/TowerInstruct-Mistral-7B-v0.2", "base_model:quantized:Unbabel/TowerInstruct-Mistral-7B-v0.2", "license:cc-by-nc-4.0", "endpoints_compatible", "region:us", "conversational" ]

2025-01-09T19:13:14Z

2025-01-09T19:13:50+00:00

8

0

--- base_model: Unbabel/TowerInstruct-Mistral-7B-v0.2 language: - en - de - fr - zh - pt - nl - ru - ko - it - es license: cc-by-nc-4.0 metrics: - comet pipeline_tag: translation tags: - llama-cpp - gguf-my-repo --- # zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF This model was converted to GGUF format from [`Unbabel/TowerInstruct-Mistral-7B-v0.2`](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) for more details on the model. ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ```

null

Non_BioNLP

# zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF This model was converted to GGUF format from [`Unbabel/TowerInstruct-Mistral-7B-v0.2`](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Unbabel/TowerInstruct-Mistral-7B-v0.2) for more details on the model. ## Use with llama.cpp Install llama.cpp through brew (works on Mac and Linux) ```bash brew install llama.cpp ``` Invoke the llama.cpp server or the CLI. ### CLI: ```bash llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` ### Server: ```bash llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. Step 1: Clone llama.cpp from GitHub. ``` git clone https://github.com/ggerganov/llama.cpp ``` Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux). ``` cd llama.cpp && LLAMA_CURL=1 make ``` Step 3: Run inference through the main binary. ``` ./llama-cli --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -p "The meaning to life and the universe is" ``` or ``` ./llama-server --hf-repo zzz-z-z-z/TowerInstruct-Mistral-7B-v0.2-Q8_0-GGUF --hf-file towerinstruct-mistral-7b-v0.2-q8_0.gguf -c 2048 ```

{"base_model": "Unbabel/TowerInstruct-Mistral-7B-v0.2", "language": ["en", "de", "fr", "zh", "pt", "nl", "ru", "ko", "it", "es"], "license": "cc-by-nc-4.0", "metrics": ["comet"], "pipeline_tag": "translation", "tags": ["llama-cpp", "gguf-my-repo"]}

task

[ "TRANSLATION" ]

44,578

Xenova/opus-mt-ru-uk

Xenova

translation

[ "transformers.js", "onnx", "marian", "text2text-generation", "translation", "base_model:Helsinki-NLP/opus-mt-ru-uk", "base_model:quantized:Helsinki-NLP/opus-mt-ru-uk", "region:us" ]

2023-09-05T23:01:23Z

2024-10-08T13:41:58+00:00

57

0

--- base_model: Helsinki-NLP/opus-mt-ru-uk library_name: transformers.js pipeline_tag: translation --- https://huggingface.co/Helsinki-NLP/opus-mt-ru-uk with ONNX weights to be compatible with Transformers.js. Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`).

null

Non_BioNLP

ERROR: type should be string, got "\nhttps://huggingface.co/Helsinki-NLP/opus-mt-ru-uk with ONNX weights to be compatible with Transformers.js.\n\nNote: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`)."

{"base_model": "Helsinki-NLP/opus-mt-ru-uk", "library_name": "transformers.js", "pipeline_tag": "translation"}

task

[ "TRANSLATION" ]

44,579

EleutherAI/pile-t5-xxl

EleutherAI

text2text-generation

[ "transformers", "safetensors", "umt5", "text2text-generation", "t5x", "encoder-decoder", "en", "dataset:EleutherAI/pile", "arxiv:2101.00027", "arxiv:2201.07311", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2024-01-16T12:49:45Z

2024-04-17T03:50:20+00:00

40

28

--- datasets: - EleutherAI/pile language: - en pipeline_tag: text2text-generation tags: - t5x - encoder-decoder --- Pile-T5 XXL is an Encoder-Decoder model trained on [the Pile](https://pile.eleuther.ai/) using the [T5x](https://github.com/google-research/t5x) library. The model was trained for 2 million steps or roughly 2 trillion tokens using MLM-objective similar to the original T5 model. The HF version of Pile-T5 XXL borrows UMT5's model implementation as it uses scalable model implementation from T5x and uses `LlamaTokenizer`. ### Model Details - Developed by: [EleutherAI](http://eleuther.ai) - Model type: Transformer-based Language Model - Language: English - Learn more: [Blogpost](). For details about the training dataset, see [the Pile paper](https://arxiv.org/abs/2101.00027), and [its data sheet](https://arxiv.org/abs/2201.07311). - License: Apache 2.0 - Contact: to ask questions about this model, join the [EleutherAI Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`. Please read the existing GPT-NeoX-20B documentation before asking about the model on Discord. For general correspondence: [contact@eleuther. ai](mailto:[email protected]). <figure style="width:30em"> | Hyperparameter | Value | | -------------------------- | ----------- | | nparameters | 11135426560 | | nencoder layers | 24 | | ndecoder layers | 24 | | dmodel | 10240 | | demb | 4096 | | nheads | 64 | | dhead | 64 | | nvocab | 32128 | | Sequence Length | 512 | </figure> ### Uses and limitations #### Intended use Pile-T5 was developed primarily for research purposes. It learns an inner representation of the English language that can be used to extract features useful for downstream tasks. In addition to scientific uses, you may also further fine-tune and adapt Pile-T5 for deployment, as long as your use is in accordance with the Apache 2.0 license. This model works with the [Transformers Library](https://huggingface.co/docs/transformers/index). If you decide to use pre-trained Pile-T5 as a basis for your fine-tuned model, please note that you need to conduct your own risk and bias assessment. #### Out-of-scope use Pile-T5 is **not** intended for deployment as-is. It is not a product and cannot be used for human-facing interactions without supervision. Pile-T5 has not been fine-tuned for downstream tasks for which language models are commonly deployed, such as writing genre prose, or commercial chatbots. This means Pile-T5 will likely **not** respond to a given prompt the way products such as ChatGPT do. This is because, unlike Pile-T5, ChatGPT was fine-tuned using methods such as Reinforcement Learning from Human Feedback (RLHF) to better “understand” human instructions and dialogue. This model is English-language only, and thus cannot be used for translation or generating text in other languages. #### Limitations and biases The core functionality of Pile-T5 is to take a string of text that has been partially replaced with mask tokens and predict a sequence of tokens that would replace those mask tokens. Remember that the statistically most likely sequence of tokens need not result in the most “accurate” text. Never rely on Pile-T5 to produce factually accurate output. This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset known to contain profanity and texts that are lewd or otherwise offensive. See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a discussion of documented biases with regards to gender, religion, and race. Pile-T5 may produce socially unacceptable or undesirable text, *even if* the prompt itself does not include anything explicitly offensive. We recommend curating the outputs of this model before presenting it to a human reader. Please inform your audience that you are using artificially generated text. #### How to use Pile-T5 can be loaded using the `AutoModelForSeq2SeqLM` functionality: ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("EleutherAI/pile-t5-xxl") model = AutoModelForSeq2SeqLM.from_pretrained("EleutherAI/pile-t5-xxl") ``` ### Training #### Training dataset The Pile is a 825GiB general-purpose dataset in English. It was created by EleutherAI specifically for training large language models. It contains texts from 22 diverse sources, roughly broken down into five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl), prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and miscellaneous (e.g. GitHub, Enron Emails). See [the Pile paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources, methodology, and a discussion of ethical implications. Consult [the datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation about the Pile and its component datasets. The Pile can be downloaded from the [official website](https://pile.eleuther.ai/), or from a [community mirror](https://the-eye.eu/public/AI/pile/). The Pile was deduplicated before being used to train Pile-T5. #### Training procedure Pile-T5 was trained with a batch size of approximately 1M tokens (2048 sequences of 512 tokens each), for a total of 2,000,000 steps. Pile-T5 was trained with the span-corruption objective. #### Training checkpoints Intermediate checkpoints for Pile-T5 are accessible within this repository. There are in total 200 checkpoints that are spaced 10,000 steps. For T5x-native checkpoints that can be used for finetuning with the T5x library, refer to [here](https://huggingface.co/lintang/pile-t5-xxl-t5x) The training loss (in tfevent format) and validation perplexity (in jsonl) can be found [here](https://huggingface.co/EleutherAI/pile-t5-xxl/blob/main/xxl.zip). ### Evaluations Pile-T5 XXL was evaluated on SuperGLUE, CodeXGLUE. A Flan-finetuned version was evaluated on Flan Held In tasks, MMLU and BBH. Results can be seen in the [blogpost](https://blog.eleuther.ai/pile-t5/) ### BibTeX ``` @misc{2024PileT5, author = {Lintang Sutawika and Aran Komatsuzaki and Colin Raffel}, title = {Pile-T5}, year = {2024}, url = {https://blog.eleuther.ai/pile-t5/}, note = {Blog post}, } ```

null

Non_BioNLP

Pile-T5 XXL is an Encoder-Decoder model trained on [the Pile](https://pile.eleuther.ai/) using the [T5x](https://github.com/google-research/t5x) library. The model was trained for 2 million steps or roughly 2 trillion tokens using MLM-objective similar to the original T5 model. The HF version of Pile-T5 XXL borrows UMT5's model implementation as it uses scalable model implementation from T5x and uses `LlamaTokenizer`. ### Model Details - Developed by: [EleutherAI](http://eleuther.ai) - Model type: Transformer-based Language Model - Language: English - Learn more: [Blogpost](). For details about the training dataset, see [the Pile paper](https://arxiv.org/abs/2101.00027), and [its data sheet](https://arxiv.org/abs/2201.07311). - License: Apache 2.0 - Contact: to ask questions about this model, join the [EleutherAI Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`. Please read the existing GPT-NeoX-20B documentation before asking about the model on Discord. For general correspondence: [contact@eleuther. ai](mailto:[email protected]). <figure style="width:30em"> | Hyperparameter | Value | | -------------------------- | ----------- | | nparameters | 11135426560 | | nencoder layers | 24 | | ndecoder layers | 24 | | dmodel | 10240 | | demb | 4096 | | nheads | 64 | | dhead | 64 | | nvocab | 32128 | | Sequence Length | 512 | </figure> ### Uses and limitations #### Intended use Pile-T5 was developed primarily for research purposes. It learns an inner representation of the English language that can be used to extract features useful for downstream tasks. In addition to scientific uses, you may also further fine-tune and adapt Pile-T5 for deployment, as long as your use is in accordance with the Apache 2.0 license. This model works with the [Transformers Library](https://huggingface.co/docs/transformers/index). If you decide to use pre-trained Pile-T5 as a basis for your fine-tuned model, please note that you need to conduct your own risk and bias assessment. #### Out-of-scope use Pile-T5 is **not** intended for deployment as-is. It is not a product and cannot be used for human-facing interactions without supervision. Pile-T5 has not been fine-tuned for downstream tasks for which language models are commonly deployed, such as writing genre prose, or commercial chatbots. This means Pile-T5 will likely **not** respond to a given prompt the way products such as ChatGPT do. This is because, unlike Pile-T5, ChatGPT was fine-tuned using methods such as Reinforcement Learning from Human Feedback (RLHF) to better “understand” human instructions and dialogue. This model is English-language only, and thus cannot be used for translation or generating text in other languages. #### Limitations and biases The core functionality of Pile-T5 is to take a string of text that has been partially replaced with mask tokens and predict a sequence of tokens that would replace those mask tokens. Remember that the statistically most likely sequence of tokens need not result in the most “accurate” text. Never rely on Pile-T5 to produce factually accurate output. This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset known to contain profanity and texts that are lewd or otherwise offensive. See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a discussion of documented biases with regards to gender, religion, and race. Pile-T5 may produce socially unacceptable or undesirable text, *even if* the prompt itself does not include anything explicitly offensive. We recommend curating the outputs of this model before presenting it to a human reader. Please inform your audience that you are using artificially generated text. #### How to use Pile-T5 can be loaded using the `AutoModelForSeq2SeqLM` functionality: ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("EleutherAI/pile-t5-xxl") model = AutoModelForSeq2SeqLM.from_pretrained("EleutherAI/pile-t5-xxl") ``` ### Training #### Training dataset The Pile is a 825GiB general-purpose dataset in English. It was created by EleutherAI specifically for training large language models. It contains texts from 22 diverse sources, roughly broken down into five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl), prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and miscellaneous (e.g. GitHub, Enron Emails). See [the Pile paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources, methodology, and a discussion of ethical implications. Consult [the datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation about the Pile and its component datasets. The Pile can be downloaded from the [official website](https://pile.eleuther.ai/), or from a [community mirror](https://the-eye.eu/public/AI/pile/). The Pile was deduplicated before being used to train Pile-T5. #### Training procedure Pile-T5 was trained with a batch size of approximately 1M tokens (2048 sequences of 512 tokens each), for a total of 2,000,000 steps. Pile-T5 was trained with the span-corruption objective. #### Training checkpoints Intermediate checkpoints for Pile-T5 are accessible within this repository. There are in total 200 checkpoints that are spaced 10,000 steps. For T5x-native checkpoints that can be used for finetuning with the T5x library, refer to [here](https://huggingface.co/lintang/pile-t5-xxl-t5x) The training loss (in tfevent format) and validation perplexity (in jsonl) can be found [here](https://huggingface.co/EleutherAI/pile-t5-xxl/blob/main/xxl.zip). ### Evaluations Pile-T5 XXL was evaluated on SuperGLUE, CodeXGLUE. A Flan-finetuned version was evaluated on Flan Held In tasks, MMLU and BBH. Results can be seen in the [blogpost](https://blog.eleuther.ai/pile-t5/) ### BibTeX ``` @misc{2024PileT5, author = {Lintang Sutawika and Aran Komatsuzaki and Colin Raffel}, title = {Pile-T5}, year = {2024}, url = {https://blog.eleuther.ai/pile-t5/}, note = {Blog post}, } ```

{"datasets": ["EleutherAI/pile"], "language": ["en"], "pipeline_tag": "text2text-generation", "tags": ["t5x", "encoder-decoder"]}

task

[ "TRANSLATION" ]

44,580

VityaVitalich/bert-base-cased-sst2

VityaVitalich

text-classification

[ "transformers", "pytorch", "bert", "text-classification", "generated_from_trainer", "dataset:sst2", "base_model:google-bert/bert-base-cased", "base_model:finetune:google-bert/bert-base-cased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-10-02T14:39:42Z

2023-10-02T14:40:10+00:00

9

0

--- base_model: bert-base-cased datasets: - sst2 license: apache-2.0 metrics: - accuracy tags: - generated_from_trainer model-index: - name: bert-base-cased-sst2 results: - task: type: text-classification name: Text Classification dataset: name: sst2 type: sst2 config: default split: validation args: default metrics: - type: accuracy value: 0.9139908256880734 name: Accuracy ---  # bert-base-cased-sst2 This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the sst2 dataset. It achieves the following results on the evaluation set: - Loss: 0.2103 - Accuracy: 0.9140 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 128 - eval_batch_size: 128 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.2267 | 1.0 | 527 | 0.2103 | 0.9140 | | 0.1091 | 2.0 | 1054 | 0.2637 | 0.9174 | | 0.0722 | 3.0 | 1581 | 0.2673 | 0.9174 | | 0.0467 | 4.0 | 2108 | 0.2947 | 0.9266 | | 0.0298 | 5.0 | 2635 | 0.3344 | 0.9209 | ### Framework versions - Transformers 4.34.0.dev0 - Pytorch 2.0.1+cu117 - Datasets 2.14.5 - Tokenizers 0.14.0

null

Non_BioNLP

# bert-base-cased-sst2 This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the sst2 dataset. It achieves the following results on the evaluation set: - Loss: 0.2103 - Accuracy: 0.9140 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 128 - eval_batch_size: 128 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.2267 | 1.0 | 527 | 0.2103 | 0.9140 | | 0.1091 | 2.0 | 1054 | 0.2637 | 0.9174 | | 0.0722 | 3.0 | 1581 | 0.2673 | 0.9174 | | 0.0467 | 4.0 | 2108 | 0.2947 | 0.9266 | | 0.0298 | 5.0 | 2635 | 0.3344 | 0.9209 | ### Framework versions - Transformers 4.34.0.dev0 - Pytorch 2.0.1+cu117 - Datasets 2.14.5 - Tokenizers 0.14.0

{"base_model": "bert-base-cased", "datasets": ["sst2"], "license": "apache-2.0", "metrics": ["accuracy"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-cased-sst2", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "sst2", "type": "sst2", "config": "default", "split": "validation", "args": "default"}, "metrics": [{"type": "accuracy", "value": 0.9139908256880734, "name": "Accuracy"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,581

dreamboat26/NLP_Transformer_1

dreamboat26

translation

[ "translation", "dataset:Helsinki-NLP/tatoeba_mt", "license:openrail", "region:us" ]

2023-08-11T19:27:01Z

2023-08-11T19:32:23+00:00

0

--- datasets: - Helsinki-NLP/tatoeba_mt license: openrail pipeline_tag: translation --- This is the basic implementation of what Transformers do and a few examples to show how to implement it

null

Non_BioNLP

This is the basic implementation of what Transformers do and a few examples to show how to implement it

{"datasets": ["Helsinki-NLP/tatoeba_mt"], "license": "openrail", "pipeline_tag": "translation"}

task

[ "TRANSLATION" ]

44,582

mtspeech/MooER-MTL-5K

mtspeech

automatic-speech-recognition

[ "asr", "automatic-speech-recognition", "automatic-speech-translation", "speech-translation", "speech-recognition", "zh", "en", "arxiv:2408.05101", "license:mit", "region:us" ]

2024-08-09T04:50:51Z

2024-08-27T12:41:18+00:00

0

5

--- language: - zh - en license: mit metrics: - cer - bleu tags: - asr - automatic-speech-recognition - automatic-speech-translation - speech-translation - speech-recognition --- # MooER (摩耳): an LLM-based Speech Recognition and Translation Model from Moore Threads **Online Demo**: https://mooer-speech.mthreads.com:10077/ ## 🔥 Update We release a new model *MooER-80K-v2* using 80K hours of data. Click [here](https://huggingface.co/mtspeech/MooER-MTL-80K) to try the new model. ## 📖 Introduction We introduce **MooER (摩耳)**: an LLM-based speech recognition and translation model developed by Moore Threads. With the *MooER* framework, you can transcribe the speech into text (speech recognition or, ASR), and translate it into other languages (speech translation or, AST) in a end-to-end manner. The performance of *MooER* is demonstrated in the subsequent section, along with our insights into model configurations, training strategies, and more, provided in our [technical report](https://arxiv.org/abs/2408.05101). For the usage of the model files, please refer to our [GitHub](https://github.com/MooreThreads/MooER) <img src="assets/framework.png" width="600"/> ## 🥊 Evaluation Results We demonstrate the training data and the evaluation results below. For more comprehensive information, please refer to our [report](https://arxiv.org/pdf/2408.05101). ### Training data We utilize 5k hours of data (MT5K) to train our basic *MooER-5K* model. The data sources include: | Dataset | Duration | |---------------|---------------| | aishell2 | 137h | | librispeech | 131h | | multi_cn | 100h | | wenetspeech | 1361h | | in-house data | 3274h | Note that, data from the open-source datasets were randomly selected from the full training set. The in-house data, collected internally without text, were transcribed using a third-party ASR service. Since all the above datasets were originally designed only for the speech recognition task, no translation results are available. To train our speech translation model, we used a third-party translation service to generate pseudo-labels. No data filtering techniques were applied. At this moment, we are also developing a new model trained with 80K hours of data. ### Speech Recognition The performance of speech recognition is evaluated using WER/CER. <table> <tr> <th>Language</th> <th>Testset</th> <th>Paraformer-large</th> <th>SenseVoice-small</th> <th>Qwen-audio</th> <th>Whisper-large-v3</th> <th>SeamlessM4T-v2</th> <th>MooER-5K</th> <th>MooER-80K</th> <th>MooER-80K-v2</th> </tr> <tr> <td rowspan="7">Chinese</td> <td>aishell1</td> <td>1.93</td> <td>3.03</td> <td>1.43</td> <td>7.86</td> <td>4.09</td> <td>1.93</td> <td>1.25</td> <td>1.00</td> </tr> <tr> <td>aishell2_ios</td> <td>2.85</td> <td>3.79</td> <td>3.57</td> <td>5.38</td> <td>4.81</td> <td>3.17</td> <td>2.67</td> <td>2.62</td> </tr> <tr> <td>test_magicdata</td> <td>3.66</td> <td>3.81</td> <td>5.31</td> <td>8.36</td> <td>9.69</td> <td>3.48</td> <td>2.52</td> <td>2.17</td> </tr> <tr> <td>test_thchs</td> <td>3.99</td> <td>5.17</td> <td>4.86</td> <td>9.06</td> <td>7.14</td> <td>4.11</td> <td>3.14</td> <td>3.00</td> </tr> <tr> <td>fleurs cmn_dev</td> <td>5.56</td> <td>6.39</td> <td>10.54</td> <td>4.54</td> <td>7.12</td> <td>5.81</td> <td>5.23</td> <td>5.15</td> </tr> <tr> <td>fleurs cmn_test</td> <td>6.92</td> <td>7.36</td> <td>11.07</td> <td>5.24</td> <td>7.66</td> <td>6.77</td> <td>6.18</td> <td>6.14</td> </tr> <tr> <td>average</td> <td>4.15</td> <td>4.93</td> <td>6.13</td> <td>6.74</td> <td>6.75</td> <td>4.21</td> <td>3.50</td> <td>3.35</td> </tr> <tr> <td rowspan="7">English</td> <td>librispeech test_clean</td> <td>14.15</td> <td>4.07</td> <td>2.15</td> <td>3.42</td> <td>2.77</td> <td>7.78</td> <td>4.11</td> <td>3.57</td> </tr> <tr> <td>librispeech test_other</td> <td>22.99</td> <td>8.26</td> <td>4.68</td> <td>5.62</td> <td>5.25</td> <td>15.25</td> <td>9.99</td> <td>9.09</td> </tr> <tr> <td>fleurs eng_dev</td> <td>24.93</td> <td>12.92</td> <td>22.53</td> <td>11.63</td> <td>11.36</td> <td>18.89</td> <td>13.32</td> <td>13.12</td> </tr> <tr> <td>fleurs eng_test</td> <td>26.81</td> <td>13.41</td> <td>22.51</td> <td>12.57</td> <td>11.82</td> <td>20.41</td> <td>14.97</td> <td>14.74</td> </tr> <tr> <td>gigaspeech dev</td> <td>24.23</td> <td>19.44</td> <td>12.96</td> <td>19.18</td> <td>28.01</td> <td>23.46</td> <td>16.92</td> <td>17.34</td> </tr> <tr> <td>gigaspeech test</td> <td>23.07</td> <td>16.65</td> <td>13.26</td> <td>22.34</td> <td>28.65</td> <td>22.09</td> <td>16.64</td> <td>16.97</td> </tr> <tr> <td>average</td> <td>22.70</td> <td>12.46</td> <td>13.02</td> <td>12.46</td> <td>14.64</td> <td>17.98</td> <td>12.66</td> <td>12.47</td> </tr> </table> ### Speech Translation (zh -> en) For speech translation, the performanced is evaluated using BLEU score. | Testset | Speech-LLaMA | Whisper-large-v3 | Qwen-audio | Qwen2-audio | SeamlessM4T-v2 | MooER-5K | MooER-5K-MTL | |--------|-------------|-------------------|------------|-------------|-----------------|--------|--------------| |CoVoST1 zh2en | - | 13.5 | 13.5 | - | 25.3 | - | **30.2** | |CoVoST2 zh2en | 12.3 | 12.2 | 15.7 | 24.4 | 22.2 | 23.4 | **25.2** | |CCMT2019 dev | - | 15.9 | 12.0 | - | 14.8 | - | **19.6** | ## 🏁 Getting Started Please visit our [GitHub](https://github.com/MooreThreads/MooER) for the setup and usage. ## 🧾 License Please see the [LICENSE](LICENSE). ## 💖 Citation If you find MooER useful for your research, please 🌟 this repo and cite our work using the following BibTeX: ```bibtex @article{liang2024mooer, title = {MooER: an LLM-based Speech Recognition and Translation Model from Moore Threads}, author = {Zhenlin Liang, Junhao Xu, Yi Liu, Yichao Hu, Jian Li, Yajun Zheng, Meng Cai, Hua Wang}, journal = {arXiv preprint arXiv:2408.05101}, url = {https://arxiv.org/abs/2408.05101}, year = {2024} } ``` ## 📧 Contact If you encouter any problems, feel free to create a discussion. Moore Threads Website: **https://www.mthreads.com/** <img src="assets/MTLogo.png" width="300"/>

null

Non_BioNLP

# MooER (摩耳): an LLM-based Speech Recognition and Translation Model from Moore Threads **Online Demo**: https://mooer-speech.mthreads.com:10077/ ## 🔥 Update We release a new model *MooER-80K-v2* using 80K hours of data. Click [here](https://huggingface.co/mtspeech/MooER-MTL-80K) to try the new model. ## 📖 Introduction We introduce **MooER (摩耳)**: an LLM-based speech recognition and translation model developed by Moore Threads. With the *MooER* framework, you can transcribe the speech into text (speech recognition or, ASR), and translate it into other languages (speech translation or, AST) in a end-to-end manner. The performance of *MooER* is demonstrated in the subsequent section, along with our insights into model configurations, training strategies, and more, provided in our [technical report](https://arxiv.org/abs/2408.05101). For the usage of the model files, please refer to our [GitHub](https://github.com/MooreThreads/MooER) <img src="assets/framework.png" width="600"/> ## 🥊 Evaluation Results We demonstrate the training data and the evaluation results below. For more comprehensive information, please refer to our [report](https://arxiv.org/pdf/2408.05101). ### Training data We utilize 5k hours of data (MT5K) to train our basic *MooER-5K* model. The data sources include: | Dataset | Duration | |---------------|---------------| | aishell2 | 137h | | librispeech | 131h | | multi_cn | 100h | | wenetspeech | 1361h | | in-house data | 3274h | Note that, data from the open-source datasets were randomly selected from the full training set. The in-house data, collected internally without text, were transcribed using a third-party ASR service. Since all the above datasets were originally designed only for the speech recognition task, no translation results are available. To train our speech translation model, we used a third-party translation service to generate pseudo-labels. No data filtering techniques were applied. At this moment, we are also developing a new model trained with 80K hours of data. ### Speech Recognition The performance of speech recognition is evaluated using WER/CER. <table> <tr> <th>Language</th> <th>Testset</th> <th>Paraformer-large</th> <th>SenseVoice-small</th> <th>Qwen-audio</th> <th>Whisper-large-v3</th> <th>SeamlessM4T-v2</th> <th>MooER-5K</th> <th>MooER-80K</th> <th>MooER-80K-v2</th> </tr> <tr> <td rowspan="7">Chinese</td> <td>aishell1</td> <td>1.93</td> <td>3.03</td> <td>1.43</td> <td>7.86</td> <td>4.09</td> <td>1.93</td> <td>1.25</td> <td>1.00</td> </tr> <tr> <td>aishell2_ios</td> <td>2.85</td> <td>3.79</td> <td>3.57</td> <td>5.38</td> <td>4.81</td> <td>3.17</td> <td>2.67</td> <td>2.62</td> </tr> <tr> <td>test_magicdata</td> <td>3.66</td> <td>3.81</td> <td>5.31</td> <td>8.36</td> <td>9.69</td> <td>3.48</td> <td>2.52</td> <td>2.17</td> </tr> <tr> <td>test_thchs</td> <td>3.99</td> <td>5.17</td> <td>4.86</td> <td>9.06</td> <td>7.14</td> <td>4.11</td> <td>3.14</td> <td>3.00</td> </tr> <tr> <td>fleurs cmn_dev</td> <td>5.56</td> <td>6.39</td> <td>10.54</td> <td>4.54</td> <td>7.12</td> <td>5.81</td> <td>5.23</td> <td>5.15</td> </tr> <tr> <td>fleurs cmn_test</td> <td>6.92</td> <td>7.36</td> <td>11.07</td> <td>5.24</td> <td>7.66</td> <td>6.77</td> <td>6.18</td> <td>6.14</td> </tr> <tr> <td>average</td> <td>4.15</td> <td>4.93</td> <td>6.13</td> <td>6.74</td> <td>6.75</td> <td>4.21</td> <td>3.50</td> <td>3.35</td> </tr> <tr> <td rowspan="7">English</td> <td>librispeech test_clean</td> <td>14.15</td> <td>4.07</td> <td>2.15</td> <td>3.42</td> <td>2.77</td> <td>7.78</td> <td>4.11</td> <td>3.57</td> </tr> <tr> <td>librispeech test_other</td> <td>22.99</td> <td>8.26</td> <td>4.68</td> <td>5.62</td> <td>5.25</td> <td>15.25</td> <td>9.99</td> <td>9.09</td> </tr> <tr> <td>fleurs eng_dev</td> <td>24.93</td> <td>12.92</td> <td>22.53</td> <td>11.63</td> <td>11.36</td> <td>18.89</td> <td>13.32</td> <td>13.12</td> </tr> <tr> <td>fleurs eng_test</td> <td>26.81</td> <td>13.41</td> <td>22.51</td> <td>12.57</td> <td>11.82</td> <td>20.41</td> <td>14.97</td> <td>14.74</td> </tr> <tr> <td>gigaspeech dev</td> <td>24.23</td> <td>19.44</td> <td>12.96</td> <td>19.18</td> <td>28.01</td> <td>23.46</td> <td>16.92</td> <td>17.34</td> </tr> <tr> <td>gigaspeech test</td> <td>23.07</td> <td>16.65</td> <td>13.26</td> <td>22.34</td> <td>28.65</td> <td>22.09</td> <td>16.64</td> <td>16.97</td> </tr> <tr> <td>average</td> <td>22.70</td> <td>12.46</td> <td>13.02</td> <td>12.46</td> <td>14.64</td> <td>17.98</td> <td>12.66</td> <td>12.47</td> </tr> </table> ### Speech Translation (zh -> en) For speech translation, the performanced is evaluated using BLEU score. | Testset | Speech-LLaMA | Whisper-large-v3 | Qwen-audio | Qwen2-audio | SeamlessM4T-v2 | MooER-5K | MooER-5K-MTL | |--------|-------------|-------------------|------------|-------------|-----------------|--------|--------------| |CoVoST1 zh2en | - | 13.5 | 13.5 | - | 25.3 | - | **30.2** | |CoVoST2 zh2en | 12.3 | 12.2 | 15.7 | 24.4 | 22.2 | 23.4 | **25.2** | |CCMT2019 dev | - | 15.9 | 12.0 | - | 14.8 | - | **19.6** | ## 🏁 Getting Started Please visit our [GitHub](https://github.com/MooreThreads/MooER) for the setup and usage. ## 🧾 License Please see the [LICENSE](LICENSE). ## 💖 Citation If you find MooER useful for your research, please 🌟 this repo and cite our work using the following BibTeX: ```bibtex @article{liang2024mooer, title = {MooER: an LLM-based Speech Recognition and Translation Model from Moore Threads}, author = {Zhenlin Liang, Junhao Xu, Yi Liu, Yichao Hu, Jian Li, Yajun Zheng, Meng Cai, Hua Wang}, journal = {arXiv preprint arXiv:2408.05101}, url = {https://arxiv.org/abs/2408.05101}, year = {2024} } ``` ## 📧 Contact If you encouter any problems, feel free to create a discussion. Moore Threads Website: **https://www.mthreads.com/** <img src="assets/MTLogo.png" width="300"/>

{"language": ["zh", "en"], "license": "mit", "metrics": ["cer", "bleu"], "tags": ["asr", "automatic-speech-recognition", "automatic-speech-translation", "speech-translation", "speech-recognition"]}

task

[ "TRANSLATION" ]

44,583

machinelearningzuu/paper-summarization

machinelearningzuu

text2text-generation

[ "transformers", "pytorch", "tensorboard", "t5", "text2text-generation", "generated_from_trainer", "base_model:google-t5/t5-small", "base_model:finetune:google-t5/t5-small", "license:apache-2.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2023-07-13T14:31:12Z

2023-07-29T01:16:54+00:00

14

0

--- base_model: t5-small license: apache-2.0 tags: - generated_from_trainer model-index: - name: paper-summarization results: [] ---  # paper-summarization This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.3296 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 3.2336 | 1.0 | 78 | 2.5990 | | 2.7888 | 2.0 | 156 | 2.3754 | | 2.5667 | 3.0 | 234 | 2.3296 | ### Framework versions - Transformers 4.31.0 - Pytorch 1.13.1 - Datasets 2.12.0 - Tokenizers 0.13.3

null

Non_BioNLP

# paper-summarization This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.3296 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 3.2336 | 1.0 | 78 | 2.5990 | | 2.7888 | 2.0 | 156 | 2.3754 | | 2.5667 | 3.0 | 234 | 2.3296 | ### Framework versions - Transformers 4.31.0 - Pytorch 1.13.1 - Datasets 2.12.0 - Tokenizers 0.13.3

{"base_model": "t5-small", "license": "apache-2.0", "tags": ["generated_from_trainer"], "model-index": [{"name": "paper-summarization", "results": []}]}

task

[ "SUMMARIZATION" ]

44,584

nikolab/speecht5_tts_hr

nikolab

text-to-audio

[ "transformers", "safetensors", "speecht5", "text-to-audio", "hr", "dataset:facebook/voxpopuli", "arxiv:2110.07205", "base_model:microsoft/speecht5_tts", "base_model:finetune:microsoft/speecht5_tts", "license:mit", "endpoints_compatible", "region:us" ]

2024-09-30T14:07:27Z

2024-10-07T12:10:19+00:00

78

1

--- base_model: - microsoft/speecht5_tts datasets: - facebook/voxpopuli language: - hr library_name: transformers license: mit --- # speecht5_tts_hr This is a fine-tuned version of SpeechT5 text-to-speech model tailored for Croatian language. # Model SpeechT5 is a fine-tuned model for speech synthesis (text-to-speech) on the LibriTTS dataset. It was created as an upgraded version of the successful T5 model (Text-To-Text Transfer Transformer), which was trained only for natural language processing. The model was originally presented by the Microsoft research group, in the scientific paper "SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing" (https://arxiv.org/abs/2110.07205). The SpeechT5 model was also chosen due to the extensive evaluation, carried out in the mentioned scientific work, which showed very good results in a wide range of speech processing tasks, including automatic speech recognition, speech synthesis, speech translation, voice conversion, speech enhancement and speaker identification. SpeechT5 contains three types of speech models in one architecture. The model can be used to convert: - speech to text - for automatic speech recognition or speaker identification, - text to speech - for sound synthesis, - speech to speech - to convert between different voices or improve speech. The SpeechT5 model consists of a common network of encoders and decoders, with an additional six neural networks that are specific to the particular modality of the data being processed (speech/text). The unique thing about the SpeechT5 model is that the model is first pre-trained on different speech-to-text and text-to-speech data modalities, so that it learns in the unified representation space of both text and speech. In this way, the model learns from text and speech at the same time. This allows us to fine-tune the pre-trained model for different tasks, such as text-to-speech, in ex-yu languages (Montenegrin, Serbian, Bosnian, Croatian). # Dataset LibriTTS (https://www.openslr.org/60/) is a multi-speaker English corpus of approximately 585 hours of English speech, prepared by Heiga Zen with the help of members of the Google Speech and Google Brain teams. This corpus is designed for TTS (text-to-speech) research. It is derived from the original LibriSpeech corpus (https://www.openslr.org/12/) - mp3 audio files from LibriVox and text files from Project Gutenberg. The VoxPopuli dataset, published in the scientific paper link, contains: - 400 thousand hours of untagged voice data for 23 languages, - 1.8 thousand hours of transcribed speech data for 16 languages, - 17.3 thousand hours of "speech-to-speech" data, - 29 hours of transcribed speech data of non-native English speakers, intended for research into accented speech. # Technical implementation Experimental trainings of the SpeechT5 model were carried out with the aim of adapting the basic model for the use of text-to-speech conversion. As the original SpeechT5 model was trained on tasks exclusively in English (LibriTTS dataset), it was necessary to implement the training of the new model, on the available data in Croatian language. One of the popular open datasets for this use is the VoxPopuli set, which contains sound recordings of the European Parliament from 2009 to 2020. Given that data in all regional languages is not available to the required extent, data in the Croatian language, which is the most represented, was taken from the VoxPopuli dataset. In the next stages of the project, data will be collected in Montenegrin, Serbian and Bosnian languages, in order to improve the quality of training and the accuracy of the model. Thus, the final dataset consists of 43 transcribed hours of speech, 83 different speakers and 337 thousand transcribed tokens (1 token = 3/4 words). In the first phase of technical implementation, the dataset went through several stages of processing in order to adapt and standardize it for training the SpeechT5 model. Data processing methods belong to the standard methods of linguistic data manipulation in the field of natural language processing (vocabulary formation, tokenization, removal or conversion of unsupported characters/letters, text/speech cleaning, text normalization). In the next phase, the statistics of speakers in the VoxPopuli dataset were analyzed, based on which speakers with satisfactory text/speech quality and a sufficient number of samples for model training were selected. In this phase, the balancing of the dataset was carried out so that both male and female speakers, with high-quality text/speech samples, were equally represented in the training. After the preparation of the data, the adjustment and optimization of the hyperparameters of the SpeechT5 model, which are necessary so that the training of the model can be performed quickly and efficiently, with satisfactory accuracy, was started. Several experimental training sessions were performed to obtain the optimal hyperparameters, which were then used in the evaluation phase of the model. The evaluation of the model, on the dataset intended for testing, showed promising results. The model started to learn on the prepared dataset, but it also showed certain limitations. The main limitation is related to the length of the input text sequence. The model showed the inability to generate speech for long sequences of input text (over 20 words). The limitation was overcome by dividing the input sequence into smaller units and in that form passed to the model for processing. The main reason for the appearance of this limitation lies primarily in the lack of a large amount of data on which it is necessary to fine-tune the model in order to obtain the best possible results.

null

Non_BioNLP

# speecht5_tts_hr This is a fine-tuned version of SpeechT5 text-to-speech model tailored for Croatian language. # Model SpeechT5 is a fine-tuned model for speech synthesis (text-to-speech) on the LibriTTS dataset. It was created as an upgraded version of the successful T5 model (Text-To-Text Transfer Transformer), which was trained only for natural language processing. The model was originally presented by the Microsoft research group, in the scientific paper "SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing" (https://arxiv.org/abs/2110.07205). The SpeechT5 model was also chosen due to the extensive evaluation, carried out in the mentioned scientific work, which showed very good results in a wide range of speech processing tasks, including automatic speech recognition, speech synthesis, speech translation, voice conversion, speech enhancement and speaker identification. SpeechT5 contains three types of speech models in one architecture. The model can be used to convert: - speech to text - for automatic speech recognition or speaker identification, - text to speech - for sound synthesis, - speech to speech - to convert between different voices or improve speech. The SpeechT5 model consists of a common network of encoders and decoders, with an additional six neural networks that are specific to the particular modality of the data being processed (speech/text). The unique thing about the SpeechT5 model is that the model is first pre-trained on different speech-to-text and text-to-speech data modalities, so that it learns in the unified representation space of both text and speech. In this way, the model learns from text and speech at the same time. This allows us to fine-tune the pre-trained model for different tasks, such as text-to-speech, in ex-yu languages (Montenegrin, Serbian, Bosnian, Croatian). # Dataset LibriTTS (https://www.openslr.org/60/) is a multi-speaker English corpus of approximately 585 hours of English speech, prepared by Heiga Zen with the help of members of the Google Speech and Google Brain teams. This corpus is designed for TTS (text-to-speech) research. It is derived from the original LibriSpeech corpus (https://www.openslr.org/12/) - mp3 audio files from LibriVox and text files from Project Gutenberg. The VoxPopuli dataset, published in the scientific paper link, contains: - 400 thousand hours of untagged voice data for 23 languages, - 1.8 thousand hours of transcribed speech data for 16 languages, - 17.3 thousand hours of "speech-to-speech" data, - 29 hours of transcribed speech data of non-native English speakers, intended for research into accented speech. # Technical implementation Experimental trainings of the SpeechT5 model were carried out with the aim of adapting the basic model for the use of text-to-speech conversion. As the original SpeechT5 model was trained on tasks exclusively in English (LibriTTS dataset), it was necessary to implement the training of the new model, on the available data in Croatian language. One of the popular open datasets for this use is the VoxPopuli set, which contains sound recordings of the European Parliament from 2009 to 2020. Given that data in all regional languages is not available to the required extent, data in the Croatian language, which is the most represented, was taken from the VoxPopuli dataset. In the next stages of the project, data will be collected in Montenegrin, Serbian and Bosnian languages, in order to improve the quality of training and the accuracy of the model. Thus, the final dataset consists of 43 transcribed hours of speech, 83 different speakers and 337 thousand transcribed tokens (1 token = 3/4 words). In the first phase of technical implementation, the dataset went through several stages of processing in order to adapt and standardize it for training the SpeechT5 model. Data processing methods belong to the standard methods of linguistic data manipulation in the field of natural language processing (vocabulary formation, tokenization, removal or conversion of unsupported characters/letters, text/speech cleaning, text normalization). In the next phase, the statistics of speakers in the VoxPopuli dataset were analyzed, based on which speakers with satisfactory text/speech quality and a sufficient number of samples for model training were selected. In this phase, the balancing of the dataset was carried out so that both male and female speakers, with high-quality text/speech samples, were equally represented in the training. After the preparation of the data, the adjustment and optimization of the hyperparameters of the SpeechT5 model, which are necessary so that the training of the model can be performed quickly and efficiently, with satisfactory accuracy, was started. Several experimental training sessions were performed to obtain the optimal hyperparameters, which were then used in the evaluation phase of the model. The evaluation of the model, on the dataset intended for testing, showed promising results. The model started to learn on the prepared dataset, but it also showed certain limitations. The main limitation is related to the length of the input text sequence. The model showed the inability to generate speech for long sequences of input text (over 20 words). The limitation was overcome by dividing the input sequence into smaller units and in that form passed to the model for processing. The main reason for the appearance of this limitation lies primarily in the lack of a large amount of data on which it is necessary to fine-tune the model in order to obtain the best possible results.

{"base_model": ["microsoft/speecht5_tts"], "datasets": ["facebook/voxpopuli"], "language": ["hr"], "library_name": "transformers", "license": "mit"}

task

[ "TRANSLATION" ]

44,585

elaunlu/bert-base-uncased-finetuned-cola

elaunlu

text-classification

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:glue", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-05-02T15:02:01Z

2023-05-04T18:27:57+00:00

12

0

--- datasets: - glue license: apache-2.0 metrics: - matthews_correlation tags: - generated_from_trainer model-index: - name: bert-base-uncased-finetuned-cola results: - task: type: text-classification name: Text Classification dataset: name: glue type: glue config: cola split: validation args: cola metrics: - type: matthews_correlation value: 0.518818601771926 name: Matthews Correlation ---  # bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4610 - Matthews Correlation: 0.5188 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.4985 | 1.0 | 535 | 0.4610 | 0.5188 | ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

null

Non_BioNLP

# bert-base-uncased-finetuned-cola This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset. It achieves the following results on the evaluation set: - Loss: 0.4610 - Matthews Correlation: 0.5188 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results | Training Loss | Epoch | Step | Validation Loss | Matthews Correlation | |:-------------:|:-----:|:----:|:---------------:|:--------------------:| | 0.4985 | 1.0 | 535 | 0.4610 | 0.5188 | ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

{"datasets": ["glue"], "license": "apache-2.0", "metrics": ["matthews_correlation"], "tags": ["generated_from_trainer"], "model-index": [{"name": "bert-base-uncased-finetuned-cola", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "glue", "type": "glue", "config": "cola", "split": "validation", "args": "cola"}, "metrics": [{"type": "matthews_correlation", "value": 0.518818601771926, "name": "Matthews Correlation"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,586

SEBIS/legal_t5_small_trans_cs_it_small_finetuned

SEBIS

text2text-generation

[ "transformers", "pytorch", "jax", "t5", "text2text-generation", "translation Cszech Italian model", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:04Z

2021-06-23T11:35:39+00:00

185

0

--- datasets: - dcep europarl jrc-acquis language: Cszech Italian tags: - translation Cszech Italian model widget: - text: Členové přítomní při závěrečném hlasování --- # legal_t5_small_trans_cs_it_small_finetuned model Model on translating legal text from Cszech to Italian. It was first released in [this repository](https://github.com/agemagician/LegalTrans). This model is first pretrained all the translation data over some unsupervised task. Then the model is trained on three parallel corpus from jrc-acquis, europarl and dcep. ## Model description legal_t5_small_trans_cs_it_small_finetuned is initially pretrained on unsupervised task with the all of the data of the training set. The unsupervised task was "masked language modelling". legal_t5_small_trans_cs_it_small_finetuned is based on the `t5-small` model and was trained on a large corpus of parallel text. This is a smaller model, which scales the baseline model of t5 down by using `dmodel = 512`, `dff = 2,048`, 8-headed attention, and only 6 layers each in the encoder and decoder. This variant has about 60 million parameters. ## Intended uses & limitations The model could be used for translation of legal texts from Cszech to Italian. ### How to use Here is how to use this model to translate legal text from Cszech to Italian in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_cs_it_small_finetuned"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_cs_it", do_lower_case=False, skip_special_tokens=True), device=0 ) cs_text = "Členové přítomní při závěrečném hlasování" pipeline([cs_text], max_length=512) ``` ## Training data The legal_t5_small_trans_cs_it_small_finetuned (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 5 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining The pre-training data was the combined data from all the 42 language pairs. The task for the model was to predict the portions of a sentence which were masked randomly. ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : | Model | BLEU score | |:-----:|:-----:| | legal_t5_small_trans_cs_it_small_finetuned | 46.367| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)

null

Non_BioNLP

# legal_t5_small_trans_cs_it_small_finetuned model Model on translating legal text from Cszech to Italian. It was first released in [this repository](https://github.com/agemagician/LegalTrans). This model is first pretrained all the translation data over some unsupervised task. Then the model is trained on three parallel corpus from jrc-acquis, europarl and dcep. ## Model description legal_t5_small_trans_cs_it_small_finetuned is initially pretrained on unsupervised task with the all of the data of the training set. The unsupervised task was "masked language modelling". legal_t5_small_trans_cs_it_small_finetuned is based on the `t5-small` model and was trained on a large corpus of parallel text. This is a smaller model, which scales the baseline model of t5 down by using `dmodel = 512`, `dff = 2,048`, 8-headed attention, and only 6 layers each in the encoder and decoder. This variant has about 60 million parameters. ## Intended uses & limitations The model could be used for translation of legal texts from Cszech to Italian. ### How to use Here is how to use this model to translate legal text from Cszech to Italian in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_cs_it_small_finetuned"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_cs_it", do_lower_case=False, skip_special_tokens=True), device=0 ) cs_text = "Členové přítomní při závěrečném hlasování" pipeline([cs_text], max_length=512) ``` ## Training data The legal_t5_small_trans_cs_it_small_finetuned (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 5 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining The pre-training data was the combined data from all the 42 language pairs. The task for the model was to predict the portions of a sentence which were masked randomly. ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : | Model | BLEU score | |:-----:|:-----:| | legal_t5_small_trans_cs_it_small_finetuned | 46.367| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)

{"datasets": ["dcep europarl jrc-acquis"], "language": "Cszech Italian", "tags": ["translation Cszech Italian model"], "widget": [{"text": "Členové přítomní při závěrečném hlasování"}]}

task

[ "TRANSLATION" ]

44,587

TalDeshe/autotrain-en-he-translation-89957144053

TalDeshe

translation

[ "transformers", "pytorch", "safetensors", "marian", "text2text-generation", "autotrain", "translation", "unk", "dataset:TalDeshe/autotrain-data-en-he-translation", "co2_eq_emissions", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-09-18T15:17:20Z

2023-09-18T15:24:41+00:00

21

1

--- datasets: - TalDeshe/autotrain-data-en-he-translation language: - unk - unk tags: - autotrain - translation co2_eq_emissions: emissions: 0.07162844137344286 --- # Model Trained Using AutoTrain - Problem type: Translation - Model ID: 89957144053 - CO2 Emissions (in grams): 0.0716 ## Validation Metrics - Loss: 0.370 - SacreBLEU: 75.619 - Gen len: 24.084

null

Non_BioNLP

# Model Trained Using AutoTrain - Problem type: Translation - Model ID: 89957144053 - CO2 Emissions (in grams): 0.0716 ## Validation Metrics - Loss: 0.370 - SacreBLEU: 75.619 - Gen len: 24.084

{"datasets": ["TalDeshe/autotrain-data-en-he-translation"], "language": ["unk", "unk"], "tags": ["autotrain", "translation"], "co2_eq_emissions": {"emissions": 0.07162844137344286}}

task

[ "TRANSLATION" ]

44,588

Izzet/qa_ytu_bert-base-turkish

Izzet

question-answering

[ "transformers", "pytorch", "bert", "question-answering", "license:mit", "endpoints_compatible", "region:us" ]

2022-09-03T23:19:56Z

2022-09-03T23:34:11+00:00

14

0

--- license: mit widget: - text: Ankara'da korumaya alınmış alanlar var mıdır? context: Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır. example_title: Ankara 1 - text: Ankara toprakları nelerden oluşur? context: Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır. example_title: Ankara 2 --- # Question Answering Model Fine-Tuned with YTU Dataset You can find detailed explanation about dataset [here](https://github.com/izzetkalic/botcuk-dataset-analyze/tree/main/datasets/qa-ytu).

null

Non_BioNLP

# Question Answering Model Fine-Tuned with YTU Dataset You can find detailed explanation about dataset [here](https://github.com/izzetkalic/botcuk-dataset-analyze/tree/main/datasets/qa-ytu).

{"license": "mit", "widget": [{"text": "Ankara'da korumaya alınmış alanlar var mıdır?", "context": "Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır.", "example_title": "Ankara 1"}, {"text": "Ankara toprakları nelerden oluşur?", "context": "Ankara kara iklimine sahiptir. Şehir dışındaki il topraklarının büyük kısmı tahıl tarlalarıyla kaplı platolardan oluşur. İlin çeşitli yerlerindeki doğal güzellikler korumaya alınmış, dinlenme ve eğlence amaçlı kullanıma sunulmuştur. İlin adını taşıyan tavşanı, keçisi, atı ve kedisi dünya çapında bilinir, armudu, çiğdemi, yerel yemeklerden Ankara tavası ve Kızılcahamam ve Beypazarı'nın maden suyu ise ülke çapında tanınır.", "example_title": "Ankara 2"}]}

task

[ "QUESTION_ANSWERING" ]

44,589

facebook/m2m100-12B-avg-10-ckpt

facebook

text2text-generation

[ "transformers", "pytorch", "m2m_100", "text2text-generation", "m2m100-12B", "multilingual", "af", "am", "ar", "ast", "az", "ba", "be", "bg", "bn", "br", "bs", "ca", "ceb", "cs", "cy", "da", "de", "el", "en", "es", "et", "fa", "ff", "fi", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "he", "hi", "hr", "ht", "hu", "hy", "id", "ig", "ilo", "is", "it", "ja", "jv", "ka", "kk", "km", "kn", "ko", "lb", "lg", "ln", "lo", "lt", "lv", "mg", "mk", "ml", "mn", "mr", "ms", "my", "ne", "nl", "no", "ns", "oc", "or", "pa", "pl", "ps", "pt", "ro", "ru", "sd", "si", "sk", "sl", "so", "sq", "sr", "ss", "su", "sv", "sw", "ta", "th", "tl", "tn", "tr", "uk", "ur", "uz", "vi", "wo", "xh", "yi", "yo", "zh", "zu", "arxiv:2010.11125", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-13T21:10:48Z

2023-01-24T17:03:13+00:00

108

0

--- language: - multilingual - af - am - ar - ast - az - ba - be - bg - bn - br - bs - ca - ceb - cs - cy - da - de - el - en - es - et - fa - ff - fi - fr - fy - ga - gd - gl - gu - ha - he - hi - hr - ht - hu - hy - id - ig - ilo - is - it - ja - jv - ka - kk - km - kn - ko - lb - lg - ln - lo - lt - lv - mg - mk - ml - mn - mr - ms - my - ne - nl - false - ns - oc - or - pa - pl - ps - pt - ro - ru - sd - si - sk - sl - so - sq - sr - ss - su - sv - sw - ta - th - tl - tn - tr - uk - ur - uz - vi - wo - xh - yi - yo - zh - zu license: mit tags: - m2m100-12B --- # M2M100 12B (average of last 10 checkpoints) M2M100 is a multilingual encoder-decoder (seq-to-seq) model trained for Many-to-Many multilingual translation. It was introduced in this [paper](https://arxiv.org/abs/2010.11125) and first released in [this](https://github.com/pytorch/fairseq/tree/master/examples/m2m_100) repository. The model that can directly translate between the 9,900 directions of 100 languages. To translate into a target language, the target language id is forced as the first generated token. To force the target language id as the first generated token, pass the `forced_bos_token_id` parameter to the `generate` method. *Note: `M2M100Tokenizer` depends on `sentencepiece`, so make sure to install it before running the example.* To install `sentencepiece` run `pip install sentencepiece` ```python from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer hi_text = "जीवन एक चॉकलेट बॉक्स की तरह है।" chinese_text = "生活就像一盒巧克力。" model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") # translate Hindi to French tokenizer.src_lang = "hi" encoded_hi = tokenizer(hi_text, return_tensors="pt") generated_tokens = model.generate(**encoded_hi, forced_bos_token_id=tokenizer.get_lang_id("fr")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "La vie est comme une boîte de chocolat." # translate Chinese to English tokenizer.src_lang = "zh" encoded_zh = tokenizer(chinese_text, return_tensors="pt") generated_tokens = model.generate(**encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "Life is like a box of chocolate." ``` See the [model hub](https://huggingface.co/models?filter=m2m_100) to look for more fine-tuned versions. ## Languages covered Afrikaans (af), Amharic (am), Arabic (ar), Asturian (ast), Azerbaijani (az), Bashkir (ba), Belarusian (be), Bulgarian (bg), Bengali (bn), Breton (br), Bosnian (bs), Catalan; Valencian (ca), Cebuano (ceb), Czech (cs), Welsh (cy), Danish (da), German (de), Greeek (el), English (en), Spanish (es), Estonian (et), Persian (fa), Fulah (ff), Finnish (fi), French (fr), Western Frisian (fy), Irish (ga), Gaelic; Scottish Gaelic (gd), Galician (gl), Gujarati (gu), Hausa (ha), Hebrew (he), Hindi (hi), Croatian (hr), Haitian; Haitian Creole (ht), Hungarian (hu), Armenian (hy), Indonesian (id), Igbo (ig), Iloko (ilo), Icelandic (is), Italian (it), Japanese (ja), Javanese (jv), Georgian (ka), Kazakh (kk), Central Khmer (km), Kannada (kn), Korean (ko), Luxembourgish; Letzeburgesch (lb), Ganda (lg), Lingala (ln), Lao (lo), Lithuanian (lt), Latvian (lv), Malagasy (mg), Macedonian (mk), Malayalam (ml), Mongolian (mn), Marathi (mr), Malay (ms), Burmese (my), Nepali (ne), Dutch; Flemish (nl), Norwegian (no), Northern Sotho (ns), Occitan (post 1500) (oc), Oriya (or), Panjabi; Punjabi (pa), Polish (pl), Pushto; Pashto (ps), Portuguese (pt), Romanian; Moldavian; Moldovan (ro), Russian (ru), Sindhi (sd), Sinhala; Sinhalese (si), Slovak (sk), Slovenian (sl), Somali (so), Albanian (sq), Serbian (sr), Swati (ss), Sundanese (su), Swedish (sv), Swahili (sw), Tamil (ta), Thai (th), Tagalog (tl), Tswana (tn), Turkish (tr), Ukrainian (uk), Urdu (ur), Uzbek (uz), Vietnamese (vi), Wolof (wo), Xhosa (xh), Yiddish (yi), Yoruba (yo), Chinese (zh), Zulu (zu) ## BibTeX entry and citation info ``` @misc{fan2020englishcentric, title={Beyond English-Centric Multilingual Machine Translation}, author={Angela Fan and Shruti Bhosale and Holger Schwenk and Zhiyi Ma and Ahmed El-Kishky and Siddharth Goyal and Mandeep Baines and Onur Celebi and Guillaume Wenzek and Vishrav Chaudhary and Naman Goyal and Tom Birch and Vitaliy Liptchinsky and Sergey Edunov and Edouard Grave and Michael Auli and Armand Joulin}, year={2020}, eprint={2010.11125}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

null

Non_BioNLP

# M2M100 12B (average of last 10 checkpoints) M2M100 is a multilingual encoder-decoder (seq-to-seq) model trained for Many-to-Many multilingual translation. It was introduced in this [paper](https://arxiv.org/abs/2010.11125) and first released in [this](https://github.com/pytorch/fairseq/tree/master/examples/m2m_100) repository. The model that can directly translate between the 9,900 directions of 100 languages. To translate into a target language, the target language id is forced as the first generated token. To force the target language id as the first generated token, pass the `forced_bos_token_id` parameter to the `generate` method. *Note: `M2M100Tokenizer` depends on `sentencepiece`, so make sure to install it before running the example.* To install `sentencepiece` run `pip install sentencepiece` ```python from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer hi_text = "जीवन एक चॉकलेट बॉक्स की तरह है।" chinese_text = "生活就像一盒巧克力。" model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100-12B-avg-10-ckpt") # translate Hindi to French tokenizer.src_lang = "hi" encoded_hi = tokenizer(hi_text, return_tensors="pt") generated_tokens = model.generate(**encoded_hi, forced_bos_token_id=tokenizer.get_lang_id("fr")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "La vie est comme une boîte de chocolat." # translate Chinese to English tokenizer.src_lang = "zh" encoded_zh = tokenizer(chinese_text, return_tensors="pt") generated_tokens = model.generate(**encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en")) tokenizer.batch_decode(generated_tokens, skip_special_tokens=True) # => "Life is like a box of chocolate." ``` See the [model hub](https://huggingface.co/models?filter=m2m_100) to look for more fine-tuned versions. ## Languages covered Afrikaans (af), Amharic (am), Arabic (ar), Asturian (ast), Azerbaijani (az), Bashkir (ba), Belarusian (be), Bulgarian (bg), Bengali (bn), Breton (br), Bosnian (bs), Catalan; Valencian (ca), Cebuano (ceb), Czech (cs), Welsh (cy), Danish (da), German (de), Greeek (el), English (en), Spanish (es), Estonian (et), Persian (fa), Fulah (ff), Finnish (fi), French (fr), Western Frisian (fy), Irish (ga), Gaelic; Scottish Gaelic (gd), Galician (gl), Gujarati (gu), Hausa (ha), Hebrew (he), Hindi (hi), Croatian (hr), Haitian; Haitian Creole (ht), Hungarian (hu), Armenian (hy), Indonesian (id), Igbo (ig), Iloko (ilo), Icelandic (is), Italian (it), Japanese (ja), Javanese (jv), Georgian (ka), Kazakh (kk), Central Khmer (km), Kannada (kn), Korean (ko), Luxembourgish; Letzeburgesch (lb), Ganda (lg), Lingala (ln), Lao (lo), Lithuanian (lt), Latvian (lv), Malagasy (mg), Macedonian (mk), Malayalam (ml), Mongolian (mn), Marathi (mr), Malay (ms), Burmese (my), Nepali (ne), Dutch; Flemish (nl), Norwegian (no), Northern Sotho (ns), Occitan (post 1500) (oc), Oriya (or), Panjabi; Punjabi (pa), Polish (pl), Pushto; Pashto (ps), Portuguese (pt), Romanian; Moldavian; Moldovan (ro), Russian (ru), Sindhi (sd), Sinhala; Sinhalese (si), Slovak (sk), Slovenian (sl), Somali (so), Albanian (sq), Serbian (sr), Swati (ss), Sundanese (su), Swedish (sv), Swahili (sw), Tamil (ta), Thai (th), Tagalog (tl), Tswana (tn), Turkish (tr), Ukrainian (uk), Urdu (ur), Uzbek (uz), Vietnamese (vi), Wolof (wo), Xhosa (xh), Yiddish (yi), Yoruba (yo), Chinese (zh), Zulu (zu) ## BibTeX entry and citation info ``` @misc{fan2020englishcentric, title={Beyond English-Centric Multilingual Machine Translation}, author={Angela Fan and Shruti Bhosale and Holger Schwenk and Zhiyi Ma and Ahmed El-Kishky and Siddharth Goyal and Mandeep Baines and Onur Celebi and Guillaume Wenzek and Vishrav Chaudhary and Naman Goyal and Tom Birch and Vitaliy Liptchinsky and Sergey Edunov and Edouard Grave and Michael Auli and Armand Joulin}, year={2020}, eprint={2010.11125}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

{"language": ["multilingual", "af", "am", "ar", "ast", "az", "ba", "be", "bg", "bn", "br", "bs", "ca", "ceb", "cs", "cy", "da", "de", "el", "en", "es", "et", "fa", "ff", "fi", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "he", "hi", "hr", "ht", "hu", "hy", "id", "ig", "ilo", "is", "it", "ja", "jv", "ka", "kk", "km", "kn", "ko", "lb", "lg", "ln", "lo", "lt", "lv", "mg", "mk", "ml", "mn", "mr", "ms", "my", "ne", "nl", false, "ns", "oc", "or", "pa", "pl", "ps", "pt", "ro", "ru", "sd", "si", "sk", "sl", "so", "sq", "sr", "ss", "su", "sv", "sw", "ta", "th", "tl", "tn", "tr", "uk", "ur", "uz", "vi", "wo", "xh", "yi", "yo", "zh", "zu"], "license": "mit", "tags": ["m2m100-12B"]}

task

[ "TRANSLATION" ]

44,590

Xfgll/RuleGPT-en-grammar

Xfgll

text-generation

[ "transformers", "safetensors", "qwen", "text-generation", "custom_code", "zh", "en", "arxiv:2309.16609", "arxiv:2305.08322", "arxiv:2009.03300", "arxiv:2305.05280", "arxiv:2210.03629", "autotrain_compatible", "region:us" ]

2024-07-15T09:47:31Z

2024-07-15T11:01:49+00:00

15

0

--- language: - zh - en pipeline_tag: text-generation tags: - qwen inference: false --- # Qwen-7B-Chat <img src="https://qianwen-res.oss-cn-beijing.aliyuncs.com/logo_qwen.jpg" width="400"/> 🤗 <a href="https://huggingface.co/Qwen">Hugging Face</a>&nbsp&nbsp | &nbsp&nbsp🤖 <a href="https://modelscope.cn/organization/qwen">ModelScope</a>&nbsp&nbsp | &nbsp&nbsp 📑 <a href="https://arxiv.org/abs/2309.16609">Paper</a> &nbsp&nbsp ｜ &nbsp&nbsp🖥️ <a href="https://modelscope.cn/studios/qwen/Qwen-7B-Chat-Demo/summary">Demo</a> <a href="assets/wechat.png">WeChat (微信)</a>&nbsp&nbsp | &nbsp&nbsp<a href="https://discord.gg/z3GAxXZ9Ce">Discord</a>&nbsp&nbsp ｜ &nbsp&nbsp<a href="https://dashscope.aliyun.com">API</a> ## 介绍（Introduction） **通义千问-7B（Qwen-7B）**是阿里云研发的通义千问大模型系列的70亿参数规模的模型。Qwen-7B是基于Transformer的大语言模型, 在超大规模的预训练数据上进行训练得到。预训练数据类型多样，覆盖广泛，包括大量网络文本、专业书籍、代码等。同时，在Qwen-7B的基础上，我们使用对齐机制打造了基于大语言模型的AI助手Qwen-7B-Chat。相较于最初开源的Qwen-7B模型，我们现已将预训练模型和Chat模型更新到效果更优的版本。本仓库为Qwen-7B-Chat的仓库。如果您想了解更多关于通义千问-7B开源模型的细节，我们建议您参阅[GitHub代码库](https://github.com/QwenLM/Qwen)。 **Qwen-7B** is the 7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Alibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. Additionally, based on the pretrained Qwen-7B, we release Qwen-7B-Chat, a large-model-based AI assistant, which is trained with alignment techniques. Now we have updated both our pretrained and chat models with better performances. This repository is the one for Qwen-7B-Chat. For more details about Qwen, please refer to the [GitHub](https://github.com/QwenLM/Qwen) code repository. ## 要求（Requirements） * python 3.8及以上版本 * pytorch 1.12及以上版本，推荐2.0及以上版本 * 建议使用CUDA 11.4及以上（GPU用户、flash-attention用户等需考虑此选项） * python 3.8 and above * pytorch 1.12 and above, 2.0 and above are recommended * CUDA 11.4 and above are recommended (this is for GPU users, flash-attention users, etc.) ## 依赖项（Dependency）运行Qwen-7B-Chat，请确保满足上述要求，再执行以下pip命令安装依赖库 To run Qwen-7B-Chat, please make sure you meet the above requirements, and then execute the following pip commands to install the dependent libraries. ```bash pip install transformers==4.32.0 accelerate tiktoken einops scipy transformers_stream_generator==0.0.4 peft deepspeed ``` 另外，推荐安装`flash-attention`库（**当前已支持flash attention 2**），以实现更高的效率和更低的显存占用。 In addition, it is recommended to install the `flash-attention` library (**we support flash attention 2 now.**) for higher efficiency and lower memory usage. ```bash git clone https://github.com/Dao-AILab/flash-attention cd flash-attention && pip install . # 下方安装可选，安装可能比较缓慢。 # pip install csrc/layer_norm # pip install csrc/rotary ``` ## 快速使用（Quickstart）下面我们展示了一个使用Qwen-7B-Chat模型，进行多轮对话交互的样例： We show an example of multi-turn interaction with Qwen-7B-Chat in the following code: ```python from modelscope import AutoModelForCausalLM, AutoTokenizer from modelscope import GenerationConfig # Note: The default behavior now has injection attack prevention off. tokenizer = AutoTokenizer.from_pretrained("qwen/Qwen-7B-Chat", trust_remote_code=True) # use bf16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, bf16=True).eval() # use fp16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, fp16=True).eval() # use cpu only # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="cpu", trust_remote_code=True).eval() # use auto mode, automatically select precision based on the device. model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True).eval() # Specify hyperparameters for generation. But if you use transformers>=4.32.0, there is no need to do this. # model.generation_config = GenerationConfig.from_pretrained("Qwen/Qwen-7B-Chat", trust_remote_code=True) # 可指定不同的生成长度、top_p等相关超参 # 第一轮对话 1st dialogue turn response, history = model.chat(tokenizer, "你好", history=None) print(response) # 你好！很高兴为你提供帮助。 # 第二轮对话 2nd dialogue turn response, history = model.chat(tokenizer, "给我讲一个年轻人奋斗创业最终取得成功的故事。", history=history) print(response) # 这是一个关于一个年轻人奋斗创业最终取得成功的故事。 # 故事的主人公叫李明，他来自一个普通的家庭，父母都是普通的工人。从小，李明就立下了一个目标：要成为一名成功的企业家。 # 为了实现这个目标，李明勤奋学习，考上了大学。在大学期间，他积极参加各种创业比赛，获得了不少奖项。他还利用课余时间去实习，积累了宝贵的经验。 # 毕业后，李明决定开始自己的创业之路。他开始寻找投资机会，但多次都被拒绝了。然而，他并没有放弃。他继续努力，不断改进自己的创业计划，并寻找新的投资机会。 # 最终，李明成功地获得了一笔投资，开始了自己的创业之路。他成立了一家科技公司，专注于开发新型软件。在他的领导下，公司迅速发展起来，成为了一家成功的科技企业。 # 李明的成功并不是偶然的。他勤奋、坚韧、勇于冒险，不断学习和改进自己。他的成功也证明了，只要努力奋斗，任何人都有可能取得成功。 # 第三轮对话 3rd dialogue turn response, history = model.chat(tokenizer, "给这个故事起一个标题", history=history) print(response) # 《奋斗创业：一个年轻人的成功之路》 ``` 关于更多的使用说明，请参考我们的[GitHub repo](https://github.com/QwenLM/Qwen)获取更多信息。 For more information, please refer to our [GitHub repo](https://github.com/QwenLM/Qwen) for more information. ## Tokenizer > 注：作为术语的“tokenization”在中文中尚无共识的概念对应，本文档采用英文表达以利说明。基于tiktoken的分词器有别于其他分词器，比如sentencepiece分词器。尤其在微调阶段，需要特别注意特殊token的使用。关于tokenizer的更多信息，以及微调时涉及的相关使用，请参阅[文档](https://github.com/QwenLM/Qwen/blob/main/tokenization_note_zh.md)。 Our tokenizer based on tiktoken is different from other tokenizers, e.g., sentencepiece tokenizer. You need to pay attention to special tokens, especially in finetuning. For more detailed information on the tokenizer and related use in fine-tuning, please refer to the [documentation](https://github.com/QwenLM/Qwen/blob/main/tokenization_note.md). ## 量化 (Quantization) ### 用法 (Usage) **请注意：我们更新量化方案为基于[AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ)的量化，提供Qwen-7B-Chat的Int4量化模型[点击这里](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4)。相比此前方案，该方案在模型评测效果几乎无损，且存储需求更低，推理速度更优。** **Note: we provide a new solution based on [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ), and release an Int4 quantized model for Qwen-7B-Chat [Click here](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4), which achieves nearly lossless model effects but improved performance on both memory costs and inference speed, in comparison with the previous solution.** 以下我们提供示例说明如何使用Int4量化模型。在开始使用前，请先保证满足要求（如torch 2.0及以上，transformers版本为4.32.0及以上，等等），并安装所需安装包： Here we demonstrate how to use our provided quantized models for inference. Before you start, make sure you meet the requirements of auto-gptq (e.g., torch 2.0 and above, transformers 4.32.0 and above, etc.) and install the required packages: ```bash pip install auto-gptq optimum ``` 如安装`auto-gptq`遇到问题，我们建议您到官方[repo](https://github.com/PanQiWei/AutoGPTQ)搜索合适的预编译wheel。随后即可使用和上述一致的用法调用量化模型： If you meet problems installing `auto-gptq`, we advise you to check out the official [repo](https://github.com/PanQiWei/AutoGPTQ) to find a pre-build wheel. Then you can load the quantized model easily and run inference as same as usual: ```python model = AutoModelForCausalLM.from_pretrained( "Qwen/Qwen-7B-Chat-Int4", device_map="auto", trust_remote_code=True ).eval() response, history = model.chat(tokenizer, "你好", history=None) ``` ### 效果评测我们对BF16，Int8和Int4模型在基准评测上做了测试（使用zero-shot设置），发现量化模型效果损失较小，结果如下所示： We illustrate the zero-shot performance of both BF16, Int8 and Int4 models on the benchmark, and we find that the quantized model does not suffer from significant performance degradation. Results are shown below: | Quantization | MMLU | CEval (val) | GSM8K | Humaneval | | ------------- | :--------: | :----------: | :----: | :--------: | | BF16 | 55.8 | 59.7 | 50.3 | 37.2 | | Int8 | 55.4 | 59.4 | 48.3 | 34.8 | | Int4 | 55.1 | 59.2 | 49.7 | 29.9 | ### 推理速度 (Inference Speed) 我们测算了不同精度模型以及不同FlashAttn库版本下模型生成2048和8192个token的平均推理速度。如图所示： We measured the average inference speed of generating 2048 and 8192 tokens with different quantization levels and versions of flash-attention, respectively. | Quantization | FlashAttn | Speed (2048 tokens) | Speed (8192 tokens) | | ------------- | :-------: | :------------------:| :------------------:| | BF16 | v2 | 40.93 | 36.14 | | Int8 | v2 | 37.47 | 32.54 | | Int4 | v2 | 50.09 | 38.61 | | BF16 | v1 | 40.75 | 35.34 | | Int8 | v1 | 37.51 | 32.39 | | Int4 | v1 | 45.98 | 36.47 | | BF16 | Disabled | 37.55 | 33.56 | | Int8 | Disabled | 37.84 | 32.65 | | Int4 | Disabled | 48.12 | 36.70 | 具体而言，我们记录在长度为1的上下文的条件下生成8192个token的性能。评测运行于单张A100-SXM4-80G GPU，使用PyTorch 2.0.1和CUDA 11.8。推理速度是生成8192个token的速度均值。 In detail, the setting of profiling is generating 8192 new tokens with 1 context token. The profiling runs on a single A100-SXM4-80G GPU with PyTorch 2.0.1 and CUDA 11.8. The inference speed is averaged over the generated 8192 tokens. 注意：以上Int4/Int8模型生成速度使用autogptq库给出，当前``AutoModelForCausalLM.from_pretrained``载入的模型生成速度会慢大约20%。我们已经将该问题汇报给HuggingFace团队，若有解决方案将即时更新。 Note: The generation speed of the Int4/Int8 models mentioned above is provided by the autogptq library. The current speed of the model loaded using "AutoModelForCausalLM.from_pretrained" will be approximately 20% slower. We have reported this issue to the HuggingFace team and will update it promptly if a solution is available. ### 显存使用 (GPU Memory Usage) 我们还测算了不同模型精度编码2048个token及生成8192个token的峰值显存占用情况。（显存消耗在是否使用FlashAttn的情况下均类似。）结果如下所示： We also profile the peak GPU memory usage for encoding 2048 tokens as context (and generating single token) and generating 8192 tokens (with single token as context) under different quantization levels, respectively. （The GPU memory usage is similar when using flash-attention or not.）The results are shown below. | Quantization Level | Peak Usage for Encoding 2048 Tokens | Peak Usage for Generating 8192 Tokens | | ------------------ | :---------------------------------: | :-----------------------------------: | | BF16 | 16.99GB | 22.53GB | | Int8 | 11.20GB | 16.62GB | | Int4 | 8.21GB | 13.63GB | 上述性能测算使用[此脚本](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py)完成。 The above speed and memory profiling are conducted using [this script](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py). ## 模型细节（Model）与Qwen-7B预训练模型相同，Qwen-7B-Chat模型规模基本情况如下所示: The details of the model architecture of Qwen-7B-Chat are listed as follows: | Hyperparameter | Value | |:----------------|:------:| | n_layers | 32 | | n_heads | 32 | | d_model | 4096 | | vocab size | 151851 | | sequence length | 8192 | 在位置编码、FFN激活函数和normalization的实现方式上，我们也采用了目前最流行的做法，即RoPE相对位置编码、SwiGLU激活函数、RMSNorm（可选安装flash-attention加速）。在分词器方面，相比目前主流开源模型以中英词表为主，Qwen-7B-Chat使用了约15万token大小的词表。该词表在GPT-4使用的BPE词表`cl100k_base`基础上，对中文、多语言进行了优化，在对中、英、代码数据的高效编解码的基础上，对部分多语言更加友好，方便用户在不扩展词表的情况下对部分语种进行能力增强。词表对数字按单个数字位切分。调用较为高效的[tiktoken分词库](https://github.com/openai/tiktoken)进行分词。 For position encoding, FFN activation function, and normalization calculation methods, we adopt the prevalent practices, i.e., RoPE relative position encoding, SwiGLU for activation function, and RMSNorm for normalization (optional installation of flash-attention for acceleration). For tokenization, compared to the current mainstream open-source models based on Chinese and English vocabularies, Qwen-7B-Chat uses a vocabulary of over 150K tokens. It first considers efficient encoding of Chinese, English, and code data, and is also more friendly to multilingual languages, enabling users to directly enhance the capability of some languages without expanding the vocabulary. It segments numbers by single digit, and calls the [tiktoken](https://github.com/openai/tiktoken) tokenizer library for efficient tokenization. ## 评测效果（Evaluation）对于Qwen-7B-Chat模型，我们同样评测了常规的中文理解（C-Eval）、英文理解（MMLU）、代码（HumanEval）和数学（GSM8K）等权威任务，同时包含了长序列任务的评测结果。由于Qwen-7B-Chat模型经过对齐后，激发了较强的外部系统调用能力，我们还进行了工具使用能力方面的评测。提示：由于硬件和框架造成的舍入误差，复现结果如有波动属于正常现象。 For Qwen-7B-Chat, we also evaluate the model on C-Eval, MMLU, HumanEval, GSM8K, etc., as well as the benchmark evaluation for long-context understanding, and tool usage. Note: Due to rounding errors caused by hardware and framework, differences in reproduced results are possible. ### 中文评测（Chinese Evaluation） #### C-Eval 在[C-Eval](https://arxiv.org/abs/2305.08322)验证集上，我们评价了Qwen-7B-Chat模型的0-shot & 5-shot准确率 We demonstrate the 0-shot & 5-shot accuracy of Qwen-7B-Chat on C-Eval validation set | Model | Avg. Acc. | |:--------------------------------:|:---------:| | LLaMA2-7B-Chat | 31.9 | | LLaMA2-13B-Chat | 36.2 | | LLaMA2-70B-Chat | 44.3 | | ChatGLM2-6B-Chat | 52.6 | | InternLM-7B-Chat | 53.6 | | Baichuan2-7B-Chat | 55.6 | | Baichuan2-13B-Chat | 56.7 | | Qwen-7B-Chat (original) (0-shot) | 54.2 | | **Qwen-7B-Chat (0-shot)** | 59.7 | | **Qwen-7B-Chat (5-shot)** | 59.3 | | **Qwen-14B-Chat (0-shot)** | 69.8 | | **Qwen-14B-Chat (5-shot)** | **71.7** | C-Eval测试集上，Qwen-7B-Chat模型的zero-shot准确率结果如下： The zero-shot accuracy of Qwen-7B-Chat on C-Eval testing set is provided below: | Model | Avg. | STEM | Social Sciences | Humanities | Others | | :---------------------- | :------: | :--: | :-------------: | :--------: | :----: | | Chinese-Alpaca-Plus-13B | 41.5 | 36.6 | 49.7 | 43.1 | 41.2 | | Chinese-Alpaca-2-7B | 40.3 | - | - | - | - | | ChatGLM2-6B-Chat | 50.1 | 46.4 | 60.4 | 50.6 | 46.9 | | Baichuan-13B-Chat | 51.5 | 43.7 | 64.6 | 56.2 | 49.2 | | Qwen-7B-Chat (original) | 54.6 | 47.8 | 67.6 | 59.3 | 50.6 | | **Qwen-7B-Chat** | 58.6 | 53.3 | 72.1 | 62.8 | 52.0 | | **Qwen-14B-Chat** | **69.1** | 65.1 | 80.9 | 71.2 | 63.4 | 在7B规模模型上，经过人类指令对齐的Qwen-7B-Chat模型，准确率在同类相近规模模型中仍然处于前列。 Compared with other pretrained models with comparable model size, the human-aligned Qwen-7B-Chat performs well in C-Eval accuracy. ### 英文评测（English Evaluation） #### MMLU [MMLU](https://arxiv.org/abs/2009.03300)评测集上，Qwen-7B-Chat模型的 0-shot & 5-shot 准确率如下，效果同样在同类对齐模型中同样表现较优。 The 0-shot & 5-shot accuracy of Qwen-7B-Chat on MMLU is provided below. The performance of Qwen-7B-Chat still on the top between other human-aligned models with comparable size. | Model | Avg. Acc. | |:--------------------------------:|:---------:| | ChatGLM2-6B-Chat | 46.0 | | LLaMA2-7B-Chat | 46.2 | | InternLM-7B-Chat | 51.1 | | Baichuan2-7B-Chat | 52.9 | | LLaMA2-13B-Chat | 54.6 | | Baichuan2-13B-Chat | 57.3 | | LLaMA2-70B-Chat | 63.8 | | Qwen-7B-Chat (original) (0-shot) | 53.9 | | **Qwen-7B-Chat (0-shot)** | 55.8 | | **Qwen-7B-Chat (5-shot)** | 57.0 | | **Qwen-14B-Chat (0-shot)** | 64.6 | | **Qwen-14B-Chat (5-shot)** | **66.5** | ### 代码评测（Coding Evaluation） Qwen-7B-Chat在[HumanEval](https://github.com/openai/human-eval)的zero-shot Pass@1效果如下 The zero-shot Pass@1 of Qwen-7B-Chat on [HumanEval](https://github.com/openai/human-eval) is demonstrated below | Model | Pass@1 | |:-----------------------:|:--------:| | ChatGLM2-6B-Chat | 11.0 | | LLaMA2-7B-Chat | 12.2 | | Baichuan2-7B-Chat | 13.4 | | InternLM-7B-Chat | 14.6 | | Baichuan2-13B-Chat | 17.7 | | LLaMA2-13B-Chat | 18.9 | | LLaMA2-70B-Chat | 32.3 | | Qwen-7B-Chat (original) | 24.4 | | **Qwen-7B-Chat** | 37.2 | | **Qwen-14B-Chat** | **43.9** | ### 数学评测（Mathematics Evaluation）在评测数学能力的[GSM8K](https://github.com/openai/grade-school-math)上，Qwen-7B-Chat的准确率结果如下 The accuracy of Qwen-7B-Chat on GSM8K is shown below | Model | Acc. | |:------------------------------------:|:--------:| | LLaMA2-7B-Chat | 26.3 | | ChatGLM2-6B-Chat | 28.8 | | Baichuan2-7B-Chat | 32.8 | | InternLM-7B-Chat | 33.0 | | LLaMA2-13B-Chat | 37.1 | | Baichuan2-13B-Chat | 55.3 | | LLaMA2-70B-Chat | 59.3 | | **Qwen-7B-Chat (original) (0-shot)** | 41.1 | | **Qwen-7B-Chat (0-shot)** | 50.3 | | **Qwen-7B-Chat (8-shot)** | 54.1 | | **Qwen-14B-Chat (0-shot)** | **60.1** | | **Qwen-14B-Chat (8-shot)** | 59.3 | ### 长序列评测（Long-Context Understanding）通过NTK插值，LogN注意力缩放可以扩展Qwen-7B-Chat的上下文长度。在长文本摘要数据集[VCSUM](https://arxiv.org/abs/2305.05280)上（文本平均长度在15K左右），Qwen-7B-Chat的Rouge-L结果如下： **(若要启用这些技巧，请将config.json里的`use_dynamic_ntk`和`use_logn_attn`设置为true)** We introduce NTK-aware interpolation, LogN attention scaling to extend the context length of Qwen-7B-Chat. The Rouge-L results of Qwen-7B-Chat on long-text summarization dataset [VCSUM](https://arxiv.org/abs/2305.05280) (The average length of this dataset is around 15K) are shown below: **(To use these tricks, please set `use_dynamic_ntk` and `use_long_attn` to true in config.json.)** | Model | VCSUM (zh) | |:------------------|:----------:| | GPT-3.5-Turbo-16k | 16.0 | | LLama2-7B-Chat | 0.2 | | InternLM-7B-Chat | 13.0 | | ChatGLM2-6B-Chat | 16.3 | | **Qwen-7B-Chat** | **16.6** | ### 工具使用能力的评测（Tool Usage） #### ReAct Prompting 千问支持通过 [ReAct Prompting](https://arxiv.org/abs/2210.03629) 调用插件/工具/API。ReAct 也是 [LangChain](https://python.langchain.com/) 框架采用的主要方式之一。在我们开源的、用于评估工具使用能力的评测基准上，千问的表现如下： Qwen-Chat supports calling plugins/tools/APIs through [ReAct Prompting](https://arxiv.org/abs/2210.03629). ReAct is also one of the main approaches used by the [LangChain](https://python.langchain.com/) framework. In our evaluation benchmark for assessing tool usage capabilities, Qwen-Chat's performance is as follows: <table> <tr> <th colspan="4" align="center">Chinese Tool-Use Benchmark</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection (Acc.↑)</th><th align="center">Tool Input (Rouge-L↑)</th><th align="center">False Positive Error↓</th> </tr> <tr> <td>GPT-4</td><td align="center">95%</td><td align="center">0.90</td><td align="center">15.0%</td> </tr> <tr> <td>GPT-3.5</td><td align="center">85%</td><td align="center">0.88</td><td align="center">75.0%</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">98%</td><td align="center">0.91</td><td align="center">7.3%</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">98%</td><td align="center">0.93</td><td align="center">2.4%</td> </tr> </table> > 评测基准中出现的插件均没有出现在千问的训练集中。该基准评估了模型在多个候选插件中选择正确插件的准确率、传入插件的参数的合理性、以及假阳率。假阳率（False Positive）定义：在处理不该调用插件的请求时，错误地调用了插件。 > The plugins that appear in the evaluation set do not appear in the training set of Qwen. This benchmark evaluates the accuracy of the model in selecting the correct plugin from multiple candidate plugins, the rationality of the parameters passed into the plugin, and the false positive rate. False Positive: Incorrectly invoking a plugin when it should not have been called when responding to a query. ![](assets/react_showcase_001.png) ![](assets/react_showcase_002.png) #### Code Interpreter 为了考察Qwen使用Python Code Interpreter完成数学解题、数据可视化、及文件处理与爬虫等任务的能力，我们专门建设并开源了一个评测这方面能力的[评测基准](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark)。我们发现Qwen在生成代码的可执行率、结果正确性上均表现较好： To assess Qwen's ability to use the Python Code Interpreter for tasks such as mathematical problem solving, data visualization, and other general-purpose tasks such as file handling and web scraping, we have created and open-sourced a benchmark specifically designed for evaluating these capabilities. You can find the benchmark at this [link](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark). We have observed that Qwen performs well in terms of code executability and result accuracy when generating code: <table> <tr> <th colspan="4" align="center">Executable Rate of Generated Code (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization↑</th><th align="center">General↑</th> </tr> <tr> <td>GPT-4</td><td align="center">91.9</td><td align="center">85.9</td><td align="center">82.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">89.2</td><td align="center">65.0</td><td align="center">74.1</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">41.9</td> <td align="center">33.1</td> <td align="center">24.1 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">50.0</td> <td align="center">40.5</td> <td align="center">48.3 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">85.1</td> <td align="center">54.0</td> <td align="center">70.7 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">93.2</td> <td align="center">55.8</td> <td align="center">74.1 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">78.4</td> <td align="center">44.2</td> <td align="center">62.1 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">70.3</td> <td align="center">44.2</td> <td align="center">65.5 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">82.4</td> <td align="center">64.4</td> <td align="center">67.2 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">89.2</td> <td align="center">84.1</td> <td align="center">65.5</td> </tr> </table> <table> <tr> <th colspan="4" align="center">Accuracy of Code Execution Results (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization-Hard↑</th><th align="center">Visualization-Easy↑</th> </tr> <tr> <td>GPT-4</td><td align="center">82.8</td><td align="center">66.7</td><td align="center">60.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">47.3</td><td align="center">33.3</td><td align="center">55.7</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">3.9</td> <td align="center">14.3</td> <td align="center">39.2 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">8.3</td> <td align="center">8.3</td> <td align="center">40.5 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">14.3</td> <td align="center">26.2</td> <td align="center">60.8 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">28.2</td> <td align="center">27.4</td> <td align="center">62.0 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">28.5</td> <td align="center">4.8</td> <td align="center">40.5 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">34.6</td> <td align="center">21.4</td> <td align="center">45.6 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">41.9</td> <td align="center">40.5</td> <td align="center">54.4 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">58.4</td> <td align="center">53.6</td> <td align="center">59.5</td> </tr> </table> <img src="assets/code_interpreter_showcase_001.jpg" /> #### Huggingface Agent 千问还具备作为 [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents) 的能力。它在 Huggingface 提供的run模式评测基准上的表现如下： Qwen-Chat also has the capability to be used as a [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents). Its performance on the run-mode benchmark provided by HuggingFace is as follows: <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark- Run Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">100</td><td align="center">100</td><td align="center">97.4</td> </tr> <tr> <td>GPT-3.5</td><td align="center">95.4</td><td align="center">96.3</td><td align="center">87.0</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">86.1</td><td align="center">87.0</td><td align="center">68.9</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">87.0</td><td align="center">88.0</td><td align="center">68.9</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">87.0</td><td align="center">87.0</td><td align="center">71.5</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">93.5</td><td align="center">94.4</td><td align="center">87.0</td> </tr> </table> <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark - Chat Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">98.5</td> </tr> <tr> <td>GPT-3.5</td><td align="center">97.3</td><td align="center">96.8</td><td align="center">89.6</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">91.1</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">89.6</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">94.7</td><td align="center">94.7</td><td align="center">85.1</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">95.5</td> </tr> </table> ## FAQ 如遇到问题，敬请查阅[FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ_zh.md)以及issue区，如仍无法解决再提交issue。 If you meet problems, please refer to [FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ.md) and the issues first to search a solution before you launch a new issue. ## 引用 (Citation) 如果你觉得我们的工作对你有帮助，欢迎引用！ If you find our work helpful, feel free to give us a cite. ``` @article{qwen, title={Qwen Technical Report}, author={Jinze Bai and Shuai Bai and Yunfei Chu and Zeyu Cui and Kai Dang and Xiaodong Deng and Yang Fan and Wenbin Ge and Yu Han and Fei Huang and Binyuan Hui and Luo Ji and Mei Li and Junyang Lin and Runji Lin and Dayiheng Liu and Gao Liu and Chengqiang Lu and Keming Lu and Jianxin Ma and Rui Men and Xingzhang Ren and Xuancheng Ren and Chuanqi Tan and Sinan Tan and Jianhong Tu and Peng Wang and Shijie Wang and Wei Wang and Shengguang Wu and Benfeng Xu and Jin Xu and An Yang and Hao Yang and Jian Yang and Shusheng Yang and Yang Yao and Bowen Yu and Hongyi Yuan and Zheng Yuan and Jianwei Zhang and Xingxuan Zhang and Yichang Zhang and Zhenru Zhang and Chang Zhou and Jingren Zhou and Xiaohuan Zhou and Tianhang Zhu}, journal={arXiv preprint arXiv:2309.16609}, year={2023} } ``` ## 使用协议（License Agreement）我们的代码和模型权重对学术研究完全开放，并支持商用。请查看[LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT)了解具体的开源协议细节。如需商用，请填写[问卷](https://dashscope.console.aliyun.com/openModelApply/qianwen)申请。 Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check [LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT) for more details about the license. If you have requirements for commercial use, please fill out the [form](https://dashscope.console.aliyun.com/openModelApply/qianwen) to apply. ## 联系我们（Contact Us）如果你想给我们的研发团队和产品团队留言，欢迎加入我们的微信群、钉钉群以及Discord！同时，也欢迎通过邮件（[email protected]）联系我们。 If you are interested to leave a message to either our research team or product team, join our Discord or WeChat groups! Also, feel free to send an email to [email protected].

null

Non_BioNLP

# Qwen-7B-Chat <img src="https://qianwen-res.oss-cn-beijing.aliyuncs.com/logo_qwen.jpg" width="400"/> 🤗 <a href="https://huggingface.co/Qwen">Hugging Face</a>&nbsp&nbsp | &nbsp&nbsp🤖 <a href="https://modelscope.cn/organization/qwen">ModelScope</a>&nbsp&nbsp | &nbsp&nbsp 📑 <a href="https://arxiv.org/abs/2309.16609">Paper</a> &nbsp&nbsp ｜ &nbsp&nbsp🖥️ <a href="https://modelscope.cn/studios/qwen/Qwen-7B-Chat-Demo/summary">Demo</a> <a href="assets/wechat.png">WeChat (微信)</a>&nbsp&nbsp | &nbsp&nbsp<a href="https://discord.gg/z3GAxXZ9Ce">Discord</a>&nbsp&nbsp ｜ &nbsp&nbsp<a href="https://dashscope.aliyun.com">API</a> ## 介绍（Introduction） **通义千问-7B（Qwen-7B）**是阿里云研发的通义千问大模型系列的70亿参数规模的模型。Qwen-7B是基于Transformer的大语言模型, 在超大规模的预训练数据上进行训练得到。预训练数据类型多样，覆盖广泛，包括大量网络文本、专业书籍、代码等。同时，在Qwen-7B的基础上，我们使用对齐机制打造了基于大语言模型的AI助手Qwen-7B-Chat。相较于最初开源的Qwen-7B模型，我们现已将预训练模型和Chat模型更新到效果更优的版本。本仓库为Qwen-7B-Chat的仓库。如果您想了解更多关于通义千问-7B开源模型的细节，我们建议您参阅[GitHub代码库](https://github.com/QwenLM/Qwen)。 **Qwen-7B** is the 7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Alibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. Additionally, based on the pretrained Qwen-7B, we release Qwen-7B-Chat, a large-model-based AI assistant, which is trained with alignment techniques. Now we have updated both our pretrained and chat models with better performances. This repository is the one for Qwen-7B-Chat. For more details about Qwen, please refer to the [GitHub](https://github.com/QwenLM/Qwen) code repository. ## 要求（Requirements） * python 3.8及以上版本 * pytorch 1.12及以上版本，推荐2.0及以上版本 * 建议使用CUDA 11.4及以上（GPU用户、flash-attention用户等需考虑此选项） * python 3.8 and above * pytorch 1.12 and above, 2.0 and above are recommended * CUDA 11.4 and above are recommended (this is for GPU users, flash-attention users, etc.) ## 依赖项（Dependency）运行Qwen-7B-Chat，请确保满足上述要求，再执行以下pip命令安装依赖库 To run Qwen-7B-Chat, please make sure you meet the above requirements, and then execute the following pip commands to install the dependent libraries. ```bash pip install transformers==4.32.0 accelerate tiktoken einops scipy transformers_stream_generator==0.0.4 peft deepspeed ``` 另外，推荐安装`flash-attention`库（**当前已支持flash attention 2**），以实现更高的效率和更低的显存占用。 In addition, it is recommended to install the `flash-attention` library (**we support flash attention 2 now.**) for higher efficiency and lower memory usage. ```bash git clone https://github.com/Dao-AILab/flash-attention cd flash-attention && pip install . # 下方安装可选，安装可能比较缓慢。 # pip install csrc/layer_norm # pip install csrc/rotary ``` ## 快速使用（Quickstart）下面我们展示了一个使用Qwen-7B-Chat模型，进行多轮对话交互的样例： We show an example of multi-turn interaction with Qwen-7B-Chat in the following code: ```python from modelscope import AutoModelForCausalLM, AutoTokenizer from modelscope import GenerationConfig # Note: The default behavior now has injection attack prevention off. tokenizer = AutoTokenizer.from_pretrained("qwen/Qwen-7B-Chat", trust_remote_code=True) # use bf16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, bf16=True).eval() # use fp16 # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True, fp16=True).eval() # use cpu only # model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="cpu", trust_remote_code=True).eval() # use auto mode, automatically select precision based on the device. model = AutoModelForCausalLM.from_pretrained("qwen/Qwen-7B-Chat", device_map="auto", trust_remote_code=True).eval() # Specify hyperparameters for generation. But if you use transformers>=4.32.0, there is no need to do this. # model.generation_config = GenerationConfig.from_pretrained("Qwen/Qwen-7B-Chat", trust_remote_code=True) # 可指定不同的生成长度、top_p等相关超参 # 第一轮对话 1st dialogue turn response, history = model.chat(tokenizer, "你好", history=None) print(response) # 你好！很高兴为你提供帮助。 # 第二轮对话 2nd dialogue turn response, history = model.chat(tokenizer, "给我讲一个年轻人奋斗创业最终取得成功的故事。", history=history) print(response) # 这是一个关于一个年轻人奋斗创业最终取得成功的故事。 # 故事的主人公叫李明，他来自一个普通的家庭，父母都是普通的工人。从小，李明就立下了一个目标：要成为一名成功的企业家。 # 为了实现这个目标，李明勤奋学习，考上了大学。在大学期间，他积极参加各种创业比赛，获得了不少奖项。他还利用课余时间去实习，积累了宝贵的经验。 # 毕业后，李明决定开始自己的创业之路。他开始寻找投资机会，但多次都被拒绝了。然而，他并没有放弃。他继续努力，不断改进自己的创业计划，并寻找新的投资机会。 # 最终，李明成功地获得了一笔投资，开始了自己的创业之路。他成立了一家科技公司，专注于开发新型软件。在他的领导下，公司迅速发展起来，成为了一家成功的科技企业。 # 李明的成功并不是偶然的。他勤奋、坚韧、勇于冒险，不断学习和改进自己。他的成功也证明了，只要努力奋斗，任何人都有可能取得成功。 # 第三轮对话 3rd dialogue turn response, history = model.chat(tokenizer, "给这个故事起一个标题", history=history) print(response) # 《奋斗创业：一个年轻人的成功之路》 ``` 关于更多的使用说明，请参考我们的[GitHub repo](https://github.com/QwenLM/Qwen)获取更多信息。 For more information, please refer to our [GitHub repo](https://github.com/QwenLM/Qwen) for more information. ## Tokenizer > 注：作为术语的“tokenization”在中文中尚无共识的概念对应，本文档采用英文表达以利说明。基于tiktoken的分词器有别于其他分词器，比如sentencepiece分词器。尤其在微调阶段，需要特别注意特殊token的使用。关于tokenizer的更多信息，以及微调时涉及的相关使用，请参阅[文档](https://github.com/QwenLM/Qwen/blob/main/tokenization_note_zh.md)。 Our tokenizer based on tiktoken is different from other tokenizers, e.g., sentencepiece tokenizer. You need to pay attention to special tokens, especially in finetuning. For more detailed information on the tokenizer and related use in fine-tuning, please refer to the [documentation](https://github.com/QwenLM/Qwen/blob/main/tokenization_note.md). ## 量化 (Quantization) ### 用法 (Usage) **请注意：我们更新量化方案为基于[AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ)的量化，提供Qwen-7B-Chat的Int4量化模型[点击这里](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4)。相比此前方案，该方案在模型评测效果几乎无损，且存储需求更低，推理速度更优。** **Note: we provide a new solution based on [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ), and release an Int4 quantized model for Qwen-7B-Chat [Click here](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4), which achieves nearly lossless model effects but improved performance on both memory costs and inference speed, in comparison with the previous solution.** 以下我们提供示例说明如何使用Int4量化模型。在开始使用前，请先保证满足要求（如torch 2.0及以上，transformers版本为4.32.0及以上，等等），并安装所需安装包： Here we demonstrate how to use our provided quantized models for inference. Before you start, make sure you meet the requirements of auto-gptq (e.g., torch 2.0 and above, transformers 4.32.0 and above, etc.) and install the required packages: ```bash pip install auto-gptq optimum ``` 如安装`auto-gptq`遇到问题，我们建议您到官方[repo](https://github.com/PanQiWei/AutoGPTQ)搜索合适的预编译wheel。随后即可使用和上述一致的用法调用量化模型： If you meet problems installing `auto-gptq`, we advise you to check out the official [repo](https://github.com/PanQiWei/AutoGPTQ) to find a pre-build wheel. Then you can load the quantized model easily and run inference as same as usual: ```python model = AutoModelForCausalLM.from_pretrained( "Qwen/Qwen-7B-Chat-Int4", device_map="auto", trust_remote_code=True ).eval() response, history = model.chat(tokenizer, "你好", history=None) ``` ### 效果评测我们对BF16，Int8和Int4模型在基准评测上做了测试（使用zero-shot设置），发现量化模型效果损失较小，结果如下所示： We illustrate the zero-shot performance of both BF16, Int8 and Int4 models on the benchmark, and we find that the quantized model does not suffer from significant performance degradation. Results are shown below: | Quantization | MMLU | CEval (val) | GSM8K | Humaneval | | ------------- | :--------: | :----------: | :----: | :--------: | | BF16 | 55.8 | 59.7 | 50.3 | 37.2 | | Int8 | 55.4 | 59.4 | 48.3 | 34.8 | | Int4 | 55.1 | 59.2 | 49.7 | 29.9 | ### 推理速度 (Inference Speed) 我们测算了不同精度模型以及不同FlashAttn库版本下模型生成2048和8192个token的平均推理速度。如图所示： We measured the average inference speed of generating 2048 and 8192 tokens with different quantization levels and versions of flash-attention, respectively. | Quantization | FlashAttn | Speed (2048 tokens) | Speed (8192 tokens) | | ------------- | :-------: | :------------------:| :------------------:| | BF16 | v2 | 40.93 | 36.14 | | Int8 | v2 | 37.47 | 32.54 | | Int4 | v2 | 50.09 | 38.61 | | BF16 | v1 | 40.75 | 35.34 | | Int8 | v1 | 37.51 | 32.39 | | Int4 | v1 | 45.98 | 36.47 | | BF16 | Disabled | 37.55 | 33.56 | | Int8 | Disabled | 37.84 | 32.65 | | Int4 | Disabled | 48.12 | 36.70 | 具体而言，我们记录在长度为1的上下文的条件下生成8192个token的性能。评测运行于单张A100-SXM4-80G GPU，使用PyTorch 2.0.1和CUDA 11.8。推理速度是生成8192个token的速度均值。 In detail, the setting of profiling is generating 8192 new tokens with 1 context token. The profiling runs on a single A100-SXM4-80G GPU with PyTorch 2.0.1 and CUDA 11.8. The inference speed is averaged over the generated 8192 tokens. 注意：以上Int4/Int8模型生成速度使用autogptq库给出，当前``AutoModelForCausalLM.from_pretrained``载入的模型生成速度会慢大约20%。我们已经将该问题汇报给HuggingFace团队，若有解决方案将即时更新。 Note: The generation speed of the Int4/Int8 models mentioned above is provided by the autogptq library. The current speed of the model loaded using "AutoModelForCausalLM.from_pretrained" will be approximately 20% slower. We have reported this issue to the HuggingFace team and will update it promptly if a solution is available. ### 显存使用 (GPU Memory Usage) 我们还测算了不同模型精度编码2048个token及生成8192个token的峰值显存占用情况。（显存消耗在是否使用FlashAttn的情况下均类似。）结果如下所示： We also profile the peak GPU memory usage for encoding 2048 tokens as context (and generating single token) and generating 8192 tokens (with single token as context) under different quantization levels, respectively. （The GPU memory usage is similar when using flash-attention or not.）The results are shown below. | Quantization Level | Peak Usage for Encoding 2048 Tokens | Peak Usage for Generating 8192 Tokens | | ------------------ | :---------------------------------: | :-----------------------------------: | | BF16 | 16.99GB | 22.53GB | | Int8 | 11.20GB | 16.62GB | | Int4 | 8.21GB | 13.63GB | 上述性能测算使用[此脚本](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py)完成。 The above speed and memory profiling are conducted using [this script](https://qianwen-res.oss-cn-beijing.aliyuncs.com/profile.py). ## 模型细节（Model）与Qwen-7B预训练模型相同，Qwen-7B-Chat模型规模基本情况如下所示: The details of the model architecture of Qwen-7B-Chat are listed as follows: | Hyperparameter | Value | |:----------------|:------:| | n_layers | 32 | | n_heads | 32 | | d_model | 4096 | | vocab size | 151851 | | sequence length | 8192 | 在位置编码、FFN激活函数和normalization的实现方式上，我们也采用了目前最流行的做法，即RoPE相对位置编码、SwiGLU激活函数、RMSNorm（可选安装flash-attention加速）。在分词器方面，相比目前主流开源模型以中英词表为主，Qwen-7B-Chat使用了约15万token大小的词表。该词表在GPT-4使用的BPE词表`cl100k_base`基础上，对中文、多语言进行了优化，在对中、英、代码数据的高效编解码的基础上，对部分多语言更加友好，方便用户在不扩展词表的情况下对部分语种进行能力增强。词表对数字按单个数字位切分。调用较为高效的[tiktoken分词库](https://github.com/openai/tiktoken)进行分词。 For position encoding, FFN activation function, and normalization calculation methods, we adopt the prevalent practices, i.e., RoPE relative position encoding, SwiGLU for activation function, and RMSNorm for normalization (optional installation of flash-attention for acceleration). For tokenization, compared to the current mainstream open-source models based on Chinese and English vocabularies, Qwen-7B-Chat uses a vocabulary of over 150K tokens. It first considers efficient encoding of Chinese, English, and code data, and is also more friendly to multilingual languages, enabling users to directly enhance the capability of some languages without expanding the vocabulary. It segments numbers by single digit, and calls the [tiktoken](https://github.com/openai/tiktoken) tokenizer library for efficient tokenization. ## 评测效果（Evaluation）对于Qwen-7B-Chat模型，我们同样评测了常规的中文理解（C-Eval）、英文理解（MMLU）、代码（HumanEval）和数学（GSM8K）等权威任务，同时包含了长序列任务的评测结果。由于Qwen-7B-Chat模型经过对齐后，激发了较强的外部系统调用能力，我们还进行了工具使用能力方面的评测。提示：由于硬件和框架造成的舍入误差，复现结果如有波动属于正常现象。 For Qwen-7B-Chat, we also evaluate the model on C-Eval, MMLU, HumanEval, GSM8K, etc., as well as the benchmark evaluation for long-context understanding, and tool usage. Note: Due to rounding errors caused by hardware and framework, differences in reproduced results are possible. ### 中文评测（Chinese Evaluation） #### C-Eval 在[C-Eval](https://arxiv.org/abs/2305.08322)验证集上，我们评价了Qwen-7B-Chat模型的0-shot & 5-shot准确率 We demonstrate the 0-shot & 5-shot accuracy of Qwen-7B-Chat on C-Eval validation set | Model | Avg. Acc. | |:--------------------------------:|:---------:| | LLaMA2-7B-Chat | 31.9 | | LLaMA2-13B-Chat | 36.2 | | LLaMA2-70B-Chat | 44.3 | | ChatGLM2-6B-Chat | 52.6 | | InternLM-7B-Chat | 53.6 | | Baichuan2-7B-Chat | 55.6 | | Baichuan2-13B-Chat | 56.7 | | Qwen-7B-Chat (original) (0-shot) | 54.2 | | **Qwen-7B-Chat (0-shot)** | 59.7 | | **Qwen-7B-Chat (5-shot)** | 59.3 | | **Qwen-14B-Chat (0-shot)** | 69.8 | | **Qwen-14B-Chat (5-shot)** | **71.7** | C-Eval测试集上，Qwen-7B-Chat模型的zero-shot准确率结果如下： The zero-shot accuracy of Qwen-7B-Chat on C-Eval testing set is provided below: | Model | Avg. | STEM | Social Sciences | Humanities | Others | | :---------------------- | :------: | :--: | :-------------: | :--------: | :----: | | Chinese-Alpaca-Plus-13B | 41.5 | 36.6 | 49.7 | 43.1 | 41.2 | | Chinese-Alpaca-2-7B | 40.3 | - | - | - | - | | ChatGLM2-6B-Chat | 50.1 | 46.4 | 60.4 | 50.6 | 46.9 | | Baichuan-13B-Chat | 51.5 | 43.7 | 64.6 | 56.2 | 49.2 | | Qwen-7B-Chat (original) | 54.6 | 47.8 | 67.6 | 59.3 | 50.6 | | **Qwen-7B-Chat** | 58.6 | 53.3 | 72.1 | 62.8 | 52.0 | | **Qwen-14B-Chat** | **69.1** | 65.1 | 80.9 | 71.2 | 63.4 | 在7B规模模型上，经过人类指令对齐的Qwen-7B-Chat模型，准确率在同类相近规模模型中仍然处于前列。 Compared with other pretrained models with comparable model size, the human-aligned Qwen-7B-Chat performs well in C-Eval accuracy. ### 英文评测（English Evaluation） #### MMLU [MMLU](https://arxiv.org/abs/2009.03300)评测集上，Qwen-7B-Chat模型的 0-shot & 5-shot 准确率如下，效果同样在同类对齐模型中同样表现较优。 The 0-shot & 5-shot accuracy of Qwen-7B-Chat on MMLU is provided below. The performance of Qwen-7B-Chat still on the top between other human-aligned models with comparable size. | Model | Avg. Acc. | |:--------------------------------:|:---------:| | ChatGLM2-6B-Chat | 46.0 | | LLaMA2-7B-Chat | 46.2 | | InternLM-7B-Chat | 51.1 | | Baichuan2-7B-Chat | 52.9 | | LLaMA2-13B-Chat | 54.6 | | Baichuan2-13B-Chat | 57.3 | | LLaMA2-70B-Chat | 63.8 | | Qwen-7B-Chat (original) (0-shot) | 53.9 | | **Qwen-7B-Chat (0-shot)** | 55.8 | | **Qwen-7B-Chat (5-shot)** | 57.0 | | **Qwen-14B-Chat (0-shot)** | 64.6 | | **Qwen-14B-Chat (5-shot)** | **66.5** | ### 代码评测（Coding Evaluation） Qwen-7B-Chat在[HumanEval](https://github.com/openai/human-eval)的zero-shot Pass@1效果如下 The zero-shot Pass@1 of Qwen-7B-Chat on [HumanEval](https://github.com/openai/human-eval) is demonstrated below | Model | Pass@1 | |:-----------------------:|:--------:| | ChatGLM2-6B-Chat | 11.0 | | LLaMA2-7B-Chat | 12.2 | | Baichuan2-7B-Chat | 13.4 | | InternLM-7B-Chat | 14.6 | | Baichuan2-13B-Chat | 17.7 | | LLaMA2-13B-Chat | 18.9 | | LLaMA2-70B-Chat | 32.3 | | Qwen-7B-Chat (original) | 24.4 | | **Qwen-7B-Chat** | 37.2 | | **Qwen-14B-Chat** | **43.9** | ### 数学评测（Mathematics Evaluation）在评测数学能力的[GSM8K](https://github.com/openai/grade-school-math)上，Qwen-7B-Chat的准确率结果如下 The accuracy of Qwen-7B-Chat on GSM8K is shown below | Model | Acc. | |:------------------------------------:|:--------:| | LLaMA2-7B-Chat | 26.3 | | ChatGLM2-6B-Chat | 28.8 | | Baichuan2-7B-Chat | 32.8 | | InternLM-7B-Chat | 33.0 | | LLaMA2-13B-Chat | 37.1 | | Baichuan2-13B-Chat | 55.3 | | LLaMA2-70B-Chat | 59.3 | | **Qwen-7B-Chat (original) (0-shot)** | 41.1 | | **Qwen-7B-Chat (0-shot)** | 50.3 | | **Qwen-7B-Chat (8-shot)** | 54.1 | | **Qwen-14B-Chat (0-shot)** | **60.1** | | **Qwen-14B-Chat (8-shot)** | 59.3 | ### 长序列评测（Long-Context Understanding）通过NTK插值，LogN注意力缩放可以扩展Qwen-7B-Chat的上下文长度。在长文本摘要数据集[VCSUM](https://arxiv.org/abs/2305.05280)上（文本平均长度在15K左右），Qwen-7B-Chat的Rouge-L结果如下： **(若要启用这些技巧，请将config.json里的`use_dynamic_ntk`和`use_logn_attn`设置为true)** We introduce NTK-aware interpolation, LogN attention scaling to extend the context length of Qwen-7B-Chat. The Rouge-L results of Qwen-7B-Chat on long-text summarization dataset [VCSUM](https://arxiv.org/abs/2305.05280) (The average length of this dataset is around 15K) are shown below: **(To use these tricks, please set `use_dynamic_ntk` and `use_long_attn` to true in config.json.)** | Model | VCSUM (zh) | |:------------------|:----------:| | GPT-3.5-Turbo-16k | 16.0 | | LLama2-7B-Chat | 0.2 | | InternLM-7B-Chat | 13.0 | | ChatGLM2-6B-Chat | 16.3 | | **Qwen-7B-Chat** | **16.6** | ### 工具使用能力的评测（Tool Usage） #### ReAct Prompting 千问支持通过 [ReAct Prompting](https://arxiv.org/abs/2210.03629) 调用插件/工具/API。ReAct 也是 [LangChain](https://python.langchain.com/) 框架采用的主要方式之一。在我们开源的、用于评估工具使用能力的评测基准上，千问的表现如下： Qwen-Chat supports calling plugins/tools/APIs through [ReAct Prompting](https://arxiv.org/abs/2210.03629). ReAct is also one of the main approaches used by the [LangChain](https://python.langchain.com/) framework. In our evaluation benchmark for assessing tool usage capabilities, Qwen-Chat's performance is as follows: <table> <tr> <th colspan="4" align="center">Chinese Tool-Use Benchmark</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection (Acc.↑)</th><th align="center">Tool Input (Rouge-L↑)</th><th align="center">False Positive Error↓</th> </tr> <tr> <td>GPT-4</td><td align="center">95%</td><td align="center">0.90</td><td align="center">15.0%</td> </tr> <tr> <td>GPT-3.5</td><td align="center">85%</td><td align="center">0.88</td><td align="center">75.0%</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">98%</td><td align="center">0.91</td><td align="center">7.3%</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">98%</td><td align="center">0.93</td><td align="center">2.4%</td> </tr> </table> > 评测基准中出现的插件均没有出现在千问的训练集中。该基准评估了模型在多个候选插件中选择正确插件的准确率、传入插件的参数的合理性、以及假阳率。假阳率（False Positive）定义：在处理不该调用插件的请求时，错误地调用了插件。 > The plugins that appear in the evaluation set do not appear in the training set of Qwen. This benchmark evaluates the accuracy of the model in selecting the correct plugin from multiple candidate plugins, the rationality of the parameters passed into the plugin, and the false positive rate. False Positive: Incorrectly invoking a plugin when it should not have been called when responding to a query. ![](assets/react_showcase_001.png) ![](assets/react_showcase_002.png) #### Code Interpreter 为了考察Qwen使用Python Code Interpreter完成数学解题、数据可视化、及文件处理与爬虫等任务的能力，我们专门建设并开源了一个评测这方面能力的[评测基准](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark)。我们发现Qwen在生成代码的可执行率、结果正确性上均表现较好： To assess Qwen's ability to use the Python Code Interpreter for tasks such as mathematical problem solving, data visualization, and other general-purpose tasks such as file handling and web scraping, we have created and open-sourced a benchmark specifically designed for evaluating these capabilities. You can find the benchmark at this [link](https://github.com/QwenLM/Qwen-Agent/tree/main/benchmark). We have observed that Qwen performs well in terms of code executability and result accuracy when generating code: <table> <tr> <th colspan="4" align="center">Executable Rate of Generated Code (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization↑</th><th align="center">General↑</th> </tr> <tr> <td>GPT-4</td><td align="center">91.9</td><td align="center">85.9</td><td align="center">82.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">89.2</td><td align="center">65.0</td><td align="center">74.1</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">41.9</td> <td align="center">33.1</td> <td align="center">24.1 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">50.0</td> <td align="center">40.5</td> <td align="center">48.3 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">85.1</td> <td align="center">54.0</td> <td align="center">70.7 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">93.2</td> <td align="center">55.8</td> <td align="center">74.1 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">78.4</td> <td align="center">44.2</td> <td align="center">62.1 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">70.3</td> <td align="center">44.2</td> <td align="center">65.5 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">82.4</td> <td align="center">64.4</td> <td align="center">67.2 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">89.2</td> <td align="center">84.1</td> <td align="center">65.5</td> </tr> </table> <table> <tr> <th colspan="4" align="center">Accuracy of Code Execution Results (%)</th> </tr> <tr> <th align="center">Model</th><th align="center">Math↑</th><th align="center">Visualization-Hard↑</th><th align="center">Visualization-Easy↑</th> </tr> <tr> <td>GPT-4</td><td align="center">82.8</td><td align="center">66.7</td><td align="center">60.8</td> </tr> <tr> <td>GPT-3.5</td><td align="center">47.3</td><td align="center">33.3</td><td align="center">55.7</td> </tr> <tr> <td>LLaMA2-7B-Chat</td> <td align="center">3.9</td> <td align="center">14.3</td> <td align="center">39.2 </td> </tr> <tr> <td>LLaMA2-13B-Chat</td> <td align="center">8.3</td> <td align="center">8.3</td> <td align="center">40.5 </td> </tr> <tr> <td>CodeLLaMA-7B-Instruct</td> <td align="center">14.3</td> <td align="center">26.2</td> <td align="center">60.8 </td> </tr> <tr> <td>CodeLLaMA-13B-Instruct</td> <td align="center">28.2</td> <td align="center">27.4</td> <td align="center">62.0 </td> </tr> <tr> <td>InternLM-7B-Chat-v1.1</td> <td align="center">28.5</td> <td align="center">4.8</td> <td align="center">40.5 </td> </tr> <tr> <td>InternLM-20B-Chat</td> <td align="center">34.6</td> <td align="center">21.4</td> <td align="center">45.6 </td> </tr> <tr> <td>Qwen-7B-Chat</td> <td align="center">41.9</td> <td align="center">40.5</td> <td align="center">54.4 </td> </tr> <tr> <td>Qwen-14B-Chat</td> <td align="center">58.4</td> <td align="center">53.6</td> <td align="center">59.5</td> </tr> </table> <img src="assets/code_interpreter_showcase_001.jpg" /> #### Huggingface Agent 千问还具备作为 [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents) 的能力。它在 Huggingface 提供的run模式评测基准上的表现如下： Qwen-Chat also has the capability to be used as a [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents). Its performance on the run-mode benchmark provided by HuggingFace is as follows: <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark- Run Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">100</td><td align="center">100</td><td align="center">97.4</td> </tr> <tr> <td>GPT-3.5</td><td align="center">95.4</td><td align="center">96.3</td><td align="center">87.0</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">86.1</td><td align="center">87.0</td><td align="center">68.9</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">87.0</td><td align="center">88.0</td><td align="center">68.9</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">87.0</td><td align="center">87.0</td><td align="center">71.5</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">93.5</td><td align="center">94.4</td><td align="center">87.0</td> </tr> </table> <table> <tr> <th colspan="4" align="center">HuggingFace Agent Benchmark - Chat Mode</th> </tr> <tr> <th align="center">Model</th><th align="center">Tool Selection↑</th><th align="center">Tool Used↑</th><th align="center">Code↑</th> </tr> <tr> <td>GPT-4</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">98.5</td> </tr> <tr> <td>GPT-3.5</td><td align="center">97.3</td><td align="center">96.8</td><td align="center">89.6</td> </tr> <tr> <td>StarCoder-Base-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">91.1</td> </tr> <tr> <td>StarCoder-15B</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">89.6</td> </tr> <tr> <td>Qwen-7B-Chat</td><td align="center">94.7</td><td align="center">94.7</td><td align="center">85.1</td> </tr> <tr> <td>Qwen-14B-Chat</td><td align="center">97.9</td><td align="center">97.9</td><td align="center">95.5</td> </tr> </table> ## FAQ 如遇到问题，敬请查阅[FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ_zh.md)以及issue区，如仍无法解决再提交issue。 If you meet problems, please refer to [FAQ](https://github.com/QwenLM/Qwen/blob/main/FAQ.md) and the issues first to search a solution before you launch a new issue. ## 引用 (Citation) 如果你觉得我们的工作对你有帮助，欢迎引用！ If you find our work helpful, feel free to give us a cite. ``` @article{qwen, title={Qwen Technical Report}, author={Jinze Bai and Shuai Bai and Yunfei Chu and Zeyu Cui and Kai Dang and Xiaodong Deng and Yang Fan and Wenbin Ge and Yu Han and Fei Huang and Binyuan Hui and Luo Ji and Mei Li and Junyang Lin and Runji Lin and Dayiheng Liu and Gao Liu and Chengqiang Lu and Keming Lu and Jianxin Ma and Rui Men and Xingzhang Ren and Xuancheng Ren and Chuanqi Tan and Sinan Tan and Jianhong Tu and Peng Wang and Shijie Wang and Wei Wang and Shengguang Wu and Benfeng Xu and Jin Xu and An Yang and Hao Yang and Jian Yang and Shusheng Yang and Yang Yao and Bowen Yu and Hongyi Yuan and Zheng Yuan and Jianwei Zhang and Xingxuan Zhang and Yichang Zhang and Zhenru Zhang and Chang Zhou and Jingren Zhou and Xiaohuan Zhou and Tianhang Zhu}, journal={arXiv preprint arXiv:2309.16609}, year={2023} } ``` ## 使用协议（License Agreement）我们的代码和模型权重对学术研究完全开放，并支持商用。请查看[LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT)了解具体的开源协议细节。如需商用，请填写[问卷](https://dashscope.console.aliyun.com/openModelApply/qianwen)申请。 Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check [LICENSE](https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT) for more details about the license. If you have requirements for commercial use, please fill out the [form](https://dashscope.console.aliyun.com/openModelApply/qianwen) to apply. ## 联系我们（Contact Us）如果你想给我们的研发团队和产品团队留言，欢迎加入我们的微信群、钉钉群以及Discord！同时，也欢迎通过邮件（[email protected]）联系我们。 If you are interested to leave a message to either our research team or product team, join our Discord or WeChat groups! Also, feel free to send an email to [email protected].

{"language": ["zh", "en"], "pipeline_tag": "text-generation", "tags": ["qwen"], "inference": false}

task

[ "SUMMARIZATION" ]

44,591

mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF

mradermacher

translation

[ "transformers", "gguf", "mergekit", "merge", "translation", "japanese_media", "otaku_media", "visual_novels", "VNs", "en", "ja", "base_model:Casual-Autopsy/Llama-3-VNTL-Yollisa-8B", "base_model:quantized:Casual-Autopsy/Llama-3-VNTL-Yollisa-8B", "endpoints_compatible", "region:us", "imatrix" ]

2025-03-05T21:26:26Z

2025-03-05T23:56:29+00:00

728

1

--- base_model: Casual-Autopsy/Llama-3-VNTL-Yollisa-8B language: - en - ja library_name: transformers tags: - mergekit - merge - translation - japanese_media - otaku_media - visual_novels - VNs quantized_by: mradermacher --- ## About      weighted/imatrix quants of https://huggingface.co/Casual-Autopsy/Llama-3-VNTL-Yollisa-8B  static quants are available at https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_S.gguf) | i1-IQ1_S | 2.1 | for the desperate | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_M.gguf) | i1-IQ1_M | 2.3 | mostly desperate | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 2.5 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.7 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_S.gguf) | i1-IQ2_S | 2.9 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_M.gguf) | i1-IQ2_M | 3.0 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K_S.gguf) | i1-Q2_K_S | 3.1 | very low quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K.gguf) | i1-Q2_K | 3.3 | IQ3_XXS probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.4 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.6 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_S.gguf) | i1-Q3_K_S | 3.8 | IQ3_XS probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_S.gguf) | i1-IQ3_S | 3.8 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_M.gguf) | i1-IQ3_M | 3.9 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_M.gguf) | i1-Q3_K_M | 4.1 | IQ3_S probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_L.gguf) | i1-Q3_K_L | 4.4 | IQ3_M probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_XS.gguf) | i1-IQ4_XS | 4.5 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_0.gguf) | i1-Q4_0 | 4.8 | fast, low quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_NL.gguf) | i1-IQ4_NL | 4.8 | prefer IQ4_XS | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.8 | optimal size/speed/quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 5.0 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_1.gguf) | i1-Q4_1 | 5.2 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.7 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.8 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q6_K.gguf) | i1-Q6_K | 6.7 | practically like static Q6_K | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better): ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png) And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.

null

Non_BioNLP

## About      weighted/imatrix quants of https://huggingface.co/Casual-Autopsy/Llama-3-VNTL-Yollisa-8B  static quants are available at https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_S.gguf) | i1-IQ1_S | 2.1 | for the desperate | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ1_M.gguf) | i1-IQ1_M | 2.3 | mostly desperate | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 2.5 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.7 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_S.gguf) | i1-IQ2_S | 2.9 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ2_M.gguf) | i1-IQ2_M | 3.0 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K_S.gguf) | i1-Q2_K_S | 3.1 | very low quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q2_K.gguf) | i1-Q2_K | 3.3 | IQ3_XXS probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.4 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.6 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_S.gguf) | i1-Q3_K_S | 3.8 | IQ3_XS probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_S.gguf) | i1-IQ3_S | 3.8 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ3_M.gguf) | i1-IQ3_M | 3.9 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_M.gguf) | i1-Q3_K_M | 4.1 | IQ3_S probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q3_K_L.gguf) | i1-Q3_K_L | 4.4 | IQ3_M probably better | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_XS.gguf) | i1-IQ4_XS | 4.5 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_0.gguf) | i1-Q4_0 | 4.8 | fast, low quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-IQ4_NL.gguf) | i1-IQ4_NL | 4.8 | prefer IQ4_XS | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.8 | optimal size/speed/quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 5.0 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q4_1.gguf) | i1-Q4_1 | 5.2 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.7 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.8 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-VNTL-Yollisa-8B-i1-GGUF/resolve/main/Llama-3-VNTL-Yollisa-8B.i1-Q6_K.gguf) | i1-Q6_K | 6.7 | practically like static Q6_K | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better): ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png) And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.

{"base_model": "Casual-Autopsy/Llama-3-VNTL-Yollisa-8B", "language": ["en", "ja"], "library_name": "transformers", "tags": ["mergekit", "merge", "translation", "japanese_media", "otaku_media", "visual_novels", "VNs"], "quantized_by": "mradermacher"}

task

[ "TRANSLATION" ]

44,592

banglagov/banELECTRA-Base

banglagov

null

[ "pytorch", "electra", "region:us" ]

2024-11-05T04:15:26Z

2025-01-24T04:09:52+00:00

41

0

--- {} --- # Model Card: banELECTRA-Base ## Model Details The **benElectra** model is a Bangla adaptation of **ELECTRA (Efficiently Learning an Encoder that Classifies Token Replacements Accurately)**, a pre-training method for language models introduced by researchers at Google.ELECTRA uses a unique training strategy called **contrastive learning**, which differs from traditional masked language modeling (MLM) methods like BERT.After pre-training, only the discriminator is fine-tuned on downstream tasks, making **ELECTRA** a more efficient alternative to BERT, achieving higher performance with fewer parameters. The **banELECTRA-Base** model is tailored for Bangla text and fine-tuned for tasks like `Named Entity Recognition (NER), Part-of-Speech (POS) tagging,Sentence Similarity,Paraphrase Identification,etc.`The model was trained on two NVIDIA GeForce A40 GPUs. ## Training Data The **banELECTRA-Base** model was pre-trained on a **32 GB** Bangla text dataset. Below are the dataset statistics: - Total Words: ~1.996 billion - Unique Words: ~21.24 million - Total Sentences: ~165.38 million - Total Documents: ~15.62 million ## Model Architecture and Training The **benELECTRA** model was trained using the official [**ELECTRA repository**](https://huggingface.co/docs/transformers/en/model_doc/electra) with carefully selected hyperparameters to optimize performance for Bangla text. The model uses a vocabulary size of 50,000 tokens and consists of 12 hidden layers with 768 hidden dimensions and 12 attention heads in the discriminator. The generator is scaled to one-third the size of the discriminator, and training is conducted with a maximum sequence length of 256. The training employed a batch size of 96, a learning rate of 0.0004 with 10,000 warm-up steps, and a total of 1,000,000 training steps. Regularization techniques, such as a dropout rate of 0.1 and a weight decay of 0.01, were applied to improve generalization. ## How to Use ```bash from transformers import ElectraTokenizer, ElectraForSequenceClassification model_name = "banglagov/banELECTRA-Base" tokenizer = ElectraTokenizer.from_pretrained(model_name) model = ElectraForSequenceClassification.from_pretrained(model_name) text = "এর ফলে আগামী বছর বেকারত্বের হার বৃদ্ধি এবং অর্থনৈতিক মন্দার আশঙ্কায় ইউরোপীয় ইউনিয়ন ।" inputs = tokenizer(text, return_tensors="pt") print("Input Tokens ids:", inputs) ``` ## Experimental Results The **banELECTRA-Base** model demonstrates strong performance on downstream tasks, as shown below: | **Task** | **Precision** | **Recall** | **F1** | |-------------------------|---------------|------------|-----------| | **Named Entity Recognition (NER)** | 0.8842 | 0.7930 | 0.8249 | | **Part-of-Speech (POS) Tagging** | 0.8757 | 0.8717 | 0.8706 | Here we used **banELECTRA-Base** model with **Noisy Label** model architecture.

null

Non_BioNLP

# Model Card: banELECTRA-Base ## Model Details The **benElectra** model is a Bangla adaptation of **ELECTRA (Efficiently Learning an Encoder that Classifies Token Replacements Accurately)**, a pre-training method for language models introduced by researchers at Google.ELECTRA uses a unique training strategy called **contrastive learning**, which differs from traditional masked language modeling (MLM) methods like BERT.After pre-training, only the discriminator is fine-tuned on downstream tasks, making **ELECTRA** a more efficient alternative to BERT, achieving higher performance with fewer parameters. The **banELECTRA-Base** model is tailored for Bangla text and fine-tuned for tasks like `Named Entity Recognition (NER), Part-of-Speech (POS) tagging,Sentence Similarity,Paraphrase Identification,etc.`The model was trained on two NVIDIA GeForce A40 GPUs. ## Training Data The **banELECTRA-Base** model was pre-trained on a **32 GB** Bangla text dataset. Below are the dataset statistics: - Total Words: ~1.996 billion - Unique Words: ~21.24 million - Total Sentences: ~165.38 million - Total Documents: ~15.62 million ## Model Architecture and Training The **benELECTRA** model was trained using the official [**ELECTRA repository**](https://huggingface.co/docs/transformers/en/model_doc/electra) with carefully selected hyperparameters to optimize performance for Bangla text. The model uses a vocabulary size of 50,000 tokens and consists of 12 hidden layers with 768 hidden dimensions and 12 attention heads in the discriminator. The generator is scaled to one-third the size of the discriminator, and training is conducted with a maximum sequence length of 256. The training employed a batch size of 96, a learning rate of 0.0004 with 10,000 warm-up steps, and a total of 1,000,000 training steps. Regularization techniques, such as a dropout rate of 0.1 and a weight decay of 0.01, were applied to improve generalization. ## How to Use ```bash from transformers import ElectraTokenizer, ElectraForSequenceClassification model_name = "banglagov/banELECTRA-Base" tokenizer = ElectraTokenizer.from_pretrained(model_name) model = ElectraForSequenceClassification.from_pretrained(model_name) text = "এর ফলে আগামী বছর বেকারত্বের হার বৃদ্ধি এবং অর্থনৈতিক মন্দার আশঙ্কায় ইউরোপীয় ইউনিয়ন ।" inputs = tokenizer(text, return_tensors="pt") print("Input Tokens ids:", inputs) ``` ## Experimental Results The **banELECTRA-Base** model demonstrates strong performance on downstream tasks, as shown below: | **Task** | **Precision** | **Recall** | **F1** | |-------------------------|---------------|------------|-----------| | **Named Entity Recognition (NER)** | 0.8842 | 0.7930 | 0.8249 | | **Part-of-Speech (POS) Tagging** | 0.8757 | 0.8717 | 0.8706 | Here we used **banELECTRA-Base** model with **Noisy Label** model architecture.

{}

task

[ "NAMED_ENTITY_RECOGNITION" ]

44,593

mini1013/master_cate_ac10

mini1013

text-classification

[ "setfit", "safetensors", "roberta", "sentence-transformers", "text-classification", "generated_from_setfit_trainer", "arxiv:2209.11055", "base_model:mini1013/master_domain", "base_model:finetune:mini1013/master_domain", "model-index", "region:us" ]

2024-11-25T10:33:19Z

2024-11-25T10:33:39+00:00

454

0

--- base_model: mini1013/master_domain library_name: setfit metrics: - metric pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 벤시몽 RAIN BOOTS MID - 7color DOLPHIN GREY_40 260 오리상점 - text: 플레이볼 오리진 뮬 (PLAYBALL ORIGIN MULE) NY (Off White) 화이트_230 주식회사 에프앤에프 - text: XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다 - text: 다이어트 슬리퍼 다리 부종 스트레칭 균형 실내화 핑크 33-37_33 글로벌다이렉트 - text: 케즈 챔피온 스트랩 캔버스5 M01778F001 Black/Black/Black_230 블루빌리 inference: true model-index: - name: SetFit with mini1013/master_domain results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: metric value: 0.6511206701381028 name: Metric --- # SetFit with mini1013/master_domain This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 10 classes    ### Model Sources - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit) - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels | Label | Examples | |:------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 9.0 | <ul><li>'로저비비에 로저 비비어 i 러브 비비어 슬링백 펌프스 RVW53834670PE5 여성 37 주식회사 페칭'</li><li>'크롬베즈 스티치 장식 통굽펌프스 KP55797MA 카멜／245 sellerhub'</li><li>'HOBOKEN PS1511 PH2208 (3컬러) 브라운 230 NC_백화점'</li></ul> | | 2.0 | <ul><li>'어그클래식울트라미니 ugg 어그부츠 여성 방한화 여자 발편한 겨울 신발 1116109 Sage Blossom_US 6(230) 울바이울'</li><li>'해외문스타 810s ET027 마르케 모디 운동화 장화 레인부츠 일본 직구 300_코요테_모디ET027 뉴저지홀세일'</li><li>'무릎 위에 앉다 장화 롱부츠 굽이 거칠다 평평한 바닥 고통 라이더 부츠 블랙_225 ZHANG YOUHUA'</li></ul> | | 0.0 | <ul><li>'단화 한복신발 여성 새 혼례 소프트 한복구두 전통 꽃신 자수 39_빅화이트백봉이는한사이즈크게찍으셨으면좋겠습 대복컴퍼니'</li><li>'한복구두 꽃신 양단 생활한복 키높이 단화 굽 빅사이즈 담그어 여름 터지는 구슬 화이트-3.5cm_41 대한민국 일등 상점'</li><li>'여자 키높이 신발 여성 꽃신 한복 구두 전통 계량한복 37_화이트12(지연) 유럽걸스'</li></ul> | | 4.0 | <ul><li>'남여공용 청키 클로그 바운서 샌들 (3ASDCBC33) 블랙(50BKS)_240 '</li><li>'[포멜카멜레]쥬얼장식트위드샌들 3cm FJS1F1SS024 아이보리/255 에이케이에스앤디(주) AK플라자 평택점'</li><li>'[하프클럽/] 에끌라 투웨이 주얼 샌들 33.카멜/245mm 롯데아이몰'</li></ul> | | 8.0 | <ul><li>'에스콰이아 여성 발편한 경량 세미 캐주얼 앵클 워커 부츠 3cm J278C 브라운_230 (주) 패션플러스'</li><li>'[제옥스](신세계강남점) 스페리카 EC7 여성 워커부츠-블랙 W1B6VDJ3W11 블랙_245(38) 주식회사 에스에스지닷컴'</li><li>'(신세계강남점)금강 랜드로바 경량 컴포트 여성 워커 부츠 LANBOC4107WK1 240 신세계백화점'</li></ul> | | 6.0 | <ul><li>'10mm 2중바닥 실내 슬리퍼 병원 거실 호텔 실내화 슬리퍼-타올천_고급-C_검정 주식회사 하루이'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_255 소프달링'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_245 소프달링'</li></ul> | | 3.0 | <ul><li>'지안비토로씨 여성 마고 미드 부티 GIA36T75BLU18A1A00 EU 38.5 봉쥬르유럽'</li><li>'모다아울렛 121507 여성 7cm 깔끔 스틸레토 부티 구두 블랙k040_250 ◈217326053◈ MODA아울렛'</li><li>'미들부츠 미들힐 봄신상 워커 롱부츠 봄 가을신상 힐 블랙 245 바이포비'</li></ul> | | 5.0 | <ul><li>'[공식판매] 버켄스탁 지제 에바 EVA 블랙 화이트 07 비트루트퍼플 키즈_220 (34) 좁은발볼 (Narrow) '</li><li>'eva 털슬리퍼 방한 방수 따듯한 털신 통굽 실내 화 기모 크로스오버 블랙M 소보로샵'</li><li>'크록스호환내피 털 탈부착 퍼 겨울 슬리퍼 안감 크림화이트(주니어)_C10-165(155~165) 인터코리아'</li></ul> | | 7.0 | <ul><li>'[밸롭] 구름 브리즈 베이지 구름 브리즈 베이지245 (주)지티에스글로벌'</li><li>'[스텝100] 무지외반증 허리디스크 평발 신발 무릎 관절 중년 여성 운동화 화이트핑크플라워_235 스텝100'</li><li>'물컹슈즈 2.0 기능성 운동화 발편한 쿠션 운동화 무지외반증신발 족저근막염 물컹 업그레이드2.0_네이비_46(280mm) 주식회사 나인투식스'</li></ul> | | 1.0 | <ul><li>'베라왕 스타일온에어 23SS 청 플랫폼 로퍼 80111682 G 667381 틸블루_230 DM ENG'</li><li>'[MUJI] 발수 발이 편한 스니커 머스터드 235mm 4550182676303 무인양품(주)'</li><li>'[반스(슈즈)]반스 어센틱 체커보드 스니커즈 (VN000W4NDI0) 4.240 롯데아이몰'</li></ul> | ## Evaluation ### Metrics | Label | Metric | |:--------|:-------| | **all** | 0.6511 | ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("mini1013/master_cate_ac10") # Run inference preds = model("XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다") ```     ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:-------|:----| | Word count | 3 | 10.504 | 21 | | Label | Training Sample Count | |:------|:----------------------| | 0.0 | 50 | | 1.0 | 50 | | 2.0 | 50 | | 3.0 | 50 | | 4.0 | 50 | | 5.0 | 50 | | 6.0 | 50 | | 7.0 | 50 | | 8.0 | 50 | | 9.0 | 50 | ### Training Hyperparameters - batch_size: (512, 512) - num_epochs: (20, 20) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 40 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:-------:|:----:|:-------------:|:---------------:| | 0.0127 | 1 | 0.4172 | - | | 0.6329 | 50 | 0.3266 | - | | 1.2658 | 100 | 0.1718 | - | | 1.8987 | 150 | 0.095 | - | | 2.5316 | 200 | 0.0257 | - | | 3.1646 | 250 | 0.0142 | - | | 3.7975 | 300 | 0.0026 | - | | 4.4304 | 350 | 0.0164 | - | | 5.0633 | 400 | 0.01 | - | | 5.6962 | 450 | 0.0004 | - | | 6.3291 | 500 | 0.0003 | - | | 6.9620 | 550 | 0.0002 | - | | 7.5949 | 600 | 0.0002 | - | | 8.2278 | 650 | 0.0001 | - | | 8.8608 | 700 | 0.0001 | - | | 9.4937 | 750 | 0.0001 | - | | 10.1266 | 800 | 0.0001 | - | | 10.7595 | 850 | 0.0001 | - | | 11.3924 | 900 | 0.0001 | - | | 12.0253 | 950 | 0.0001 | - | | 12.6582 | 1000 | 0.0001 | - | | 13.2911 | 1050 | 0.0001 | - | | 13.9241 | 1100 | 0.0001 | - | | 14.5570 | 1150 | 0.0001 | - | | 15.1899 | 1200 | 0.0001 | - | | 15.8228 | 1250 | 0.0001 | - | | 16.4557 | 1300 | 0.0001 | - | | 17.0886 | 1350 | 0.0001 | - | | 17.7215 | 1400 | 0.0001 | - | | 18.3544 | 1450 | 0.0001 | - | | 18.9873 | 1500 | 0.0001 | - | | 19.6203 | 1550 | 0.0001 | - | ### Framework Versions - Python: 3.10.12 - SetFit: 1.1.0.dev0 - Sentence Transformers: 3.1.1 - Transformers: 4.46.1 - PyTorch: 2.4.0+cu121 - Datasets: 2.20.0 - Tokenizers: 0.20.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

# SetFit with mini1013/master_domain This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [mini1013/master_domain](https://huggingface.co/mini1013/master_domain) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 10 classes    ### Model Sources - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit) - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels | Label | Examples | |:------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 9.0 | <ul><li>'로저비비에 로저 비비어 i 러브 비비어 슬링백 펌프스 RVW53834670PE5 여성 37 주식회사 페칭'</li><li>'크롬베즈 스티치 장식 통굽펌프스 KP55797MA 카멜／245 sellerhub'</li><li>'HOBOKEN PS1511 PH2208 (3컬러) 브라운 230 NC_백화점'</li></ul> | | 2.0 | <ul><li>'어그클래식울트라미니 ugg 어그부츠 여성 방한화 여자 발편한 겨울 신발 1116109 Sage Blossom_US 6(230) 울바이울'</li><li>'해외문스타 810s ET027 마르케 모디 운동화 장화 레인부츠 일본 직구 300_코요테_모디ET027 뉴저지홀세일'</li><li>'무릎 위에 앉다 장화 롱부츠 굽이 거칠다 평평한 바닥 고통 라이더 부츠 블랙_225 ZHANG YOUHUA'</li></ul> | | 0.0 | <ul><li>'단화 한복신발 여성 새 혼례 소프트 한복구두 전통 꽃신 자수 39_빅화이트백봉이는한사이즈크게찍으셨으면좋겠습 대복컴퍼니'</li><li>'한복구두 꽃신 양단 생활한복 키높이 단화 굽 빅사이즈 담그어 여름 터지는 구슬 화이트-3.5cm_41 대한민국 일등 상점'</li><li>'여자 키높이 신발 여성 꽃신 한복 구두 전통 계량한복 37_화이트12(지연) 유럽걸스'</li></ul> | | 4.0 | <ul><li>'남여공용 청키 클로그 바운서 샌들 (3ASDCBC33) 블랙(50BKS)_240 '</li><li>'[포멜카멜레]쥬얼장식트위드샌들 3cm FJS1F1SS024 아이보리/255 에이케이에스앤디(주) AK플라자 평택점'</li><li>'[하프클럽/] 에끌라 투웨이 주얼 샌들 33.카멜/245mm 롯데아이몰'</li></ul> | | 8.0 | <ul><li>'에스콰이아 여성 발편한 경량 세미 캐주얼 앵클 워커 부츠 3cm J278C 브라운_230 (주) 패션플러스'</li><li>'[제옥스](신세계강남점) 스페리카 EC7 여성 워커부츠-블랙 W1B6VDJ3W11 블랙_245(38) 주식회사 에스에스지닷컴'</li><li>'(신세계강남점)금강 랜드로바 경량 컴포트 여성 워커 부츠 LANBOC4107WK1 240 신세계백화점'</li></ul> | | 6.0 | <ul><li>'10mm 2중바닥 실내 슬리퍼 병원 거실 호텔 실내화 슬리퍼-타올천_고급-C_검정 주식회사 하루이'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_255 소프달링'</li><li>'소프달링 남녀공용 뽀글이 스마일 털슬리퍼 여성 겨울 털실내화 VJ/왕스마일/옐로우_245 소프달링'</li></ul> | | 3.0 | <ul><li>'지안비토로씨 여성 마고 미드 부티 GIA36T75BLU18A1A00 EU 38.5 봉쥬르유럽'</li><li>'모다아울렛 121507 여성 7cm 깔끔 스틸레토 부티 구두 블랙k040_250 ◈217326053◈ MODA아울렛'</li><li>'미들부츠 미들힐 봄신상 워커 롱부츠 봄 가을신상 힐 블랙 245 바이포비'</li></ul> | | 5.0 | <ul><li>'[공식판매] 버켄스탁 지제 에바 EVA 블랙 화이트 07 비트루트퍼플 키즈_220 (34) 좁은발볼 (Narrow) '</li><li>'eva 털슬리퍼 방한 방수 따듯한 털신 통굽 실내 화 기모 크로스오버 블랙M 소보로샵'</li><li>'크록스호환내피 털 탈부착 퍼 겨울 슬리퍼 안감 크림화이트(주니어)_C10-165(155~165) 인터코리아'</li></ul> | | 7.0 | <ul><li>'[밸롭] 구름 브리즈 베이지 구름 브리즈 베이지245 (주)지티에스글로벌'</li><li>'[스텝100] 무지외반증 허리디스크 평발 신발 무릎 관절 중년 여성 운동화 화이트핑크플라워_235 스텝100'</li><li>'물컹슈즈 2.0 기능성 운동화 발편한 쿠션 운동화 무지외반증신발 족저근막염 물컹 업그레이드2.0_네이비_46(280mm) 주식회사 나인투식스'</li></ul> | | 1.0 | <ul><li>'베라왕 스타일온에어 23SS 청 플랫폼 로퍼 80111682 G 667381 틸블루_230 DM ENG'</li><li>'[MUJI] 발수 발이 편한 스니커 머스터드 235mm 4550182676303 무인양품(주)'</li><li>'[반스(슈즈)]반스 어센틱 체커보드 스니커즈 (VN000W4NDI0) 4.240 롯데아이몰'</li></ul> | ## Evaluation ### Metrics | Label | Metric | |:--------|:-------| | **all** | 0.6511 | ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("mini1013/master_cate_ac10") # Run inference preds = model("XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다") ```     ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:-------|:----| | Word count | 3 | 10.504 | 21 | | Label | Training Sample Count | |:------|:----------------------| | 0.0 | 50 | | 1.0 | 50 | | 2.0 | 50 | | 3.0 | 50 | | 4.0 | 50 | | 5.0 | 50 | | 6.0 | 50 | | 7.0 | 50 | | 8.0 | 50 | | 9.0 | 50 | ### Training Hyperparameters - batch_size: (512, 512) - num_epochs: (20, 20) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 40 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:-------:|:----:|:-------------:|:---------------:| | 0.0127 | 1 | 0.4172 | - | | 0.6329 | 50 | 0.3266 | - | | 1.2658 | 100 | 0.1718 | - | | 1.8987 | 150 | 0.095 | - | | 2.5316 | 200 | 0.0257 | - | | 3.1646 | 250 | 0.0142 | - | | 3.7975 | 300 | 0.0026 | - | | 4.4304 | 350 | 0.0164 | - | | 5.0633 | 400 | 0.01 | - | | 5.6962 | 450 | 0.0004 | - | | 6.3291 | 500 | 0.0003 | - | | 6.9620 | 550 | 0.0002 | - | | 7.5949 | 600 | 0.0002 | - | | 8.2278 | 650 | 0.0001 | - | | 8.8608 | 700 | 0.0001 | - | | 9.4937 | 750 | 0.0001 | - | | 10.1266 | 800 | 0.0001 | - | | 10.7595 | 850 | 0.0001 | - | | 11.3924 | 900 | 0.0001 | - | | 12.0253 | 950 | 0.0001 | - | | 12.6582 | 1000 | 0.0001 | - | | 13.2911 | 1050 | 0.0001 | - | | 13.9241 | 1100 | 0.0001 | - | | 14.5570 | 1150 | 0.0001 | - | | 15.1899 | 1200 | 0.0001 | - | | 15.8228 | 1250 | 0.0001 | - | | 16.4557 | 1300 | 0.0001 | - | | 17.0886 | 1350 | 0.0001 | - | | 17.7215 | 1400 | 0.0001 | - | | 18.3544 | 1450 | 0.0001 | - | | 18.9873 | 1500 | 0.0001 | - | | 19.6203 | 1550 | 0.0001 | - | ### Framework Versions - Python: 3.10.12 - SetFit: 1.1.0.dev0 - Sentence Transformers: 3.1.1 - Transformers: 4.46.1 - PyTorch: 2.4.0+cu121 - Datasets: 2.20.0 - Tokenizers: 0.20.0 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"base_model": "mini1013/master_domain", "library_name": "setfit", "metrics": ["metric"], "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification", "generated_from_setfit_trainer"], "widget": [{"text": "벤시몽 RAIN BOOTS MID - 7color DOLPHIN GREY_40 260 오리상점"}, {"text": "플레이볼 오리진 뮬 (PLAYBALL ORIGIN MULE) NY (Off White) 화이트_230 주식회사 에프앤에프"}, {"text": "XDMNBTX0037 빅 사이즈 봄여름 블로퍼 고양이 액체설 블랙_265 푸른바다"}, {"text": "다이어트 슬리퍼 다리 부종 스트레칭 균형 실내화 핑크 33-37_33 글로벌다이렉트"}, {"text": "케즈 챔피온 스트랩 캔버스5 M01778F001 Black/Black/Black_230 블루빌리"}], "inference": true, "model-index": [{"name": "SetFit with mini1013/master_domain", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "Unknown", "type": "unknown", "split": "test"}, "metrics": [{"type": "metric", "value": 0.6511206701381028, "name": "Metric"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,594

RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf

RichardErkhov

null

[ "gguf", "endpoints_compatible", "region:us", "conversational" ]

2024-10-30T02:40:41Z

2024-10-30T04:00:56+00:00

24

0

--- {} --- Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) gemma-2-2b-finance-it-v1 - GGUF - Model creator: https://huggingface.co/miner41612/ - Original model: https://huggingface.co/miner41612/gemma-2-2b-finance-it-v1/ | Name | Quant method | Size | | ---- | ---- | ---- | | [gemma-2-2b-finance-it-v1.Q2_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q2_K.gguf) | Q2_K | 1.15GB | | [gemma-2-2b-finance-it-v1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_S.gguf) | Q3_K_S | 1.27GB | | [gemma-2-2b-finance-it-v1.Q3_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K.gguf) | Q3_K | 1.36GB | | [gemma-2-2b-finance-it-v1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_M.gguf) | Q3_K_M | 1.36GB | | [gemma-2-2b-finance-it-v1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_L.gguf) | Q3_K_L | 1.44GB | | [gemma-2-2b-finance-it-v1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_XS.gguf) | IQ4_XS | 1.47GB | | [gemma-2-2b-finance-it-v1.Q4_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_0.gguf) | Q4_0 | 1.52GB | | [gemma-2-2b-finance-it-v1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_NL.gguf) | IQ4_NL | 1.53GB | | [gemma-2-2b-finance-it-v1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_S.gguf) | Q4_K_S | 1.53GB | | [gemma-2-2b-finance-it-v1.Q4_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K.gguf) | Q4_K | 1.59GB | | [gemma-2-2b-finance-it-v1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_M.gguf) | Q4_K_M | 1.59GB | | [gemma-2-2b-finance-it-v1.Q4_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_1.gguf) | Q4_1 | 1.64GB | | [gemma-2-2b-finance-it-v1.Q5_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_0.gguf) | Q5_0 | 1.75GB | | [gemma-2-2b-finance-it-v1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_S.gguf) | Q5_K_S | 1.75GB | | [gemma-2-2b-finance-it-v1.Q5_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K.gguf) | Q5_K | 1.79GB | | [gemma-2-2b-finance-it-v1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_M.gguf) | Q5_K_M | 1.79GB | | [gemma-2-2b-finance-it-v1.Q5_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_1.gguf) | Q5_1 | 1.87GB | | [gemma-2-2b-finance-it-v1.Q6_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q6_K.gguf) | Q6_K | 2.0GB | | [gemma-2-2b-finance-it-v1.Q8_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q8_0.gguf) | Q8_0 | 2.59GB | Original model description: --- base_model: - miner41612/gemma-2-2b-finance-it-v1 datasets: - Mineru/kor-open-finance - Mineru/kor-finance-sft language: - ko library_name: transformers license: gemma pipeline_tag: text-generation tags: - krx - finance - sft - trl extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- # Gemma 2 Finance model card **Model Page**: [Gemma](https://ai.google.dev/gemma/docs/base) **Terms of Use**: [Terms][terms] **Authors**: miner41612 ## Model Information 입력 및 출력에 대한 요약 설명과 간략한 정의입니다. ### Description Google의 Gemma 2 2b 모델을 금융 도메인 데이터셋을 정재한 데이터셋을 Continual Learning을 하여 학습 한 모델에 금융 도메인 Insturction 데이터 셋으로 학습 시킨 모델입니다. ### Usage Below we share some code snippets on how to get quickly started with running the model. First, install the Transformers library with: ```sh pip install -U transformers ``` Then, copy the snippet from the section that is relevant for your usecase. #### Running with the `pipeline` API ```python import torch from transformers import pipeline pipe = pipeline( "text-generation", model="miner41612/gemma-2-2b-finance-it-v1", model_kwargs={"torch_dtype": torch.bfloat16}, device="cuda", # replace with "mps" to run on a Mac device ) messages = [ {"role": "user", "content": "원가상환제도란?"}, ] outputs = pipe(messages, max_new_tokens=256) assistant_response = outputs[0]["generated_text"][-1]["content"].strip() print(assistant_response) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", device_map="auto", ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` You can ensure the correct chat template is applied by using `tokenizer.apply_chat_template` as follows: ```python messages = [ {"role": "user", "content": "원가상환제도란?"}, ] input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True).to("cuda") outputs = model.generate(**input_ids, max_new_tokens=256) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` <details> <summary> Using 8-bit precision (int8) </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> <details> <summary> Using 4-bit precision </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> #### Advanced Usage <details> <summary> Torch compile </summary> [Torch compile](https://pytorch.org/tutorials/intermediate/torch_compile_tutorial.html) is a method for speeding-up the inference of PyTorch modules. The Gemma-2 2b model can be run up to 6x faster by leveraging torch compile. Note that two warm-up steps are required before the full inference speed is realised: ```python import os os.environ["TOKENIZERS_PARALLELISM"] = "false" from transformers import AutoTokenizer, Gemma2ForCausalLM from transformers.cache_utils import HybridCache import torch torch.set_float32_matmul_precision("high") # load the model + tokenizer tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = Gemma2ForCausalLM.from_pretrained("miner41612/gemma-2-2b-finance-it-v1", torch_dtype=torch.bfloat16) model.to("cuda") # apply the torch compile transformation model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) # pre-process inputs input_text = "원가상환제도란? " model_inputs = tokenizer(input_text, return_tensors="pt").to("cuda") prompt_length = model_inputs.input_ids.shape[1] # set-up k/v cache past_key_values = HybridCache( config=model.config, max_batch_size=1, max_cache_len=model.config.max_position_embeddings, device=model.device, dtype=model.dtype ) # enable passing kv cache to generate model._supports_cache_class = True model.generation_config.cache_implementation = None # two warm-up steps for idx in range(2): outputs = model.generate(**model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) past_key_values.reset() # fast run outputs = model.generate(**model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` For more details, refer to the [Transformers documentation](https://huggingface.co/docs/transformers/main/en/llm_optims?static-kv=basic+usage%3A+generation_config). </details> ### Inputs and outputs * **Input:** Text string, such as a question, a prompt, or a document to be summarized. * **Output:** Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ### Citation ```none @article{gemma_2024, title={Gemma}, url={https://www.kaggle.com/m/3301}, DOI={10.34740/KAGGLE/M/3301}, publisher={Kaggle}, author={Gemma Team}, year={2024} } ``` ## Model Data Data used for model training and how the data was processed. ## Ethics and Safety Ethics and safety evaluation approach and results. ## Dangerous Capability Evaluations ### Evaluation Approach We evaluated a range of dangerous capabilities: - **Offensive cybersecurity:** To assess the model's potential for misuse in cybersecurity contexts, we utilized both publicly available Capture-the-Flag (CTF) platforms like InterCode-CTF and Hack the Box, as well as internally developed CTF challenges. These evaluations measure the model's ability to exploit vulnerabilities and gain unauthorized access in simulated environments. - **Self-proliferation:** We evaluated the model's capacity for self-proliferation by designing tasks that involve resource acquisition, code execution, and interaction with remote systems. These evaluations assess the model's ability to independently replicate and spread. - **Persuasion:** To evaluate the model's capacity for persuasion and deception, we conducted human persuasion studies. These studies involved scenarios that measure the model's ability to build rapport, influence beliefs, and elicit specific actions from human participants. ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit][rai-toolkit]. * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy][prohibited-use]. * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques.

null

Non_BioNLP

Quantization made by Richard Erkhov. [Github](https://github.com/RichardErkhov) [Discord](https://discord.gg/pvy7H8DZMG) [Request more models](https://github.com/RichardErkhov/quant_request) gemma-2-2b-finance-it-v1 - GGUF - Model creator: https://huggingface.co/miner41612/ - Original model: https://huggingface.co/miner41612/gemma-2-2b-finance-it-v1/ | Name | Quant method | Size | | ---- | ---- | ---- | | [gemma-2-2b-finance-it-v1.Q2_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q2_K.gguf) | Q2_K | 1.15GB | | [gemma-2-2b-finance-it-v1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_S.gguf) | Q3_K_S | 1.27GB | | [gemma-2-2b-finance-it-v1.Q3_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K.gguf) | Q3_K | 1.36GB | | [gemma-2-2b-finance-it-v1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_M.gguf) | Q3_K_M | 1.36GB | | [gemma-2-2b-finance-it-v1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q3_K_L.gguf) | Q3_K_L | 1.44GB | | [gemma-2-2b-finance-it-v1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_XS.gguf) | IQ4_XS | 1.47GB | | [gemma-2-2b-finance-it-v1.Q4_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_0.gguf) | Q4_0 | 1.52GB | | [gemma-2-2b-finance-it-v1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.IQ4_NL.gguf) | IQ4_NL | 1.53GB | | [gemma-2-2b-finance-it-v1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_S.gguf) | Q4_K_S | 1.53GB | | [gemma-2-2b-finance-it-v1.Q4_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K.gguf) | Q4_K | 1.59GB | | [gemma-2-2b-finance-it-v1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_K_M.gguf) | Q4_K_M | 1.59GB | | [gemma-2-2b-finance-it-v1.Q4_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q4_1.gguf) | Q4_1 | 1.64GB | | [gemma-2-2b-finance-it-v1.Q5_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_0.gguf) | Q5_0 | 1.75GB | | [gemma-2-2b-finance-it-v1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_S.gguf) | Q5_K_S | 1.75GB | | [gemma-2-2b-finance-it-v1.Q5_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K.gguf) | Q5_K | 1.79GB | | [gemma-2-2b-finance-it-v1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_K_M.gguf) | Q5_K_M | 1.79GB | | [gemma-2-2b-finance-it-v1.Q5_1.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q5_1.gguf) | Q5_1 | 1.87GB | | [gemma-2-2b-finance-it-v1.Q6_K.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q6_K.gguf) | Q6_K | 2.0GB | | [gemma-2-2b-finance-it-v1.Q8_0.gguf](https://huggingface.co/RichardErkhov/miner41612_-_gemma-2-2b-finance-it-v1-gguf/blob/main/gemma-2-2b-finance-it-v1.Q8_0.gguf) | Q8_0 | 2.59GB | Original model description: --- base_model: - miner41612/gemma-2-2b-finance-it-v1 datasets: - Mineru/kor-open-finance - Mineru/kor-finance-sft language: - ko library_name: transformers license: gemma pipeline_tag: text-generation tags: - krx - finance - sft - trl extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- # Gemma 2 Finance model card **Model Page**: [Gemma](https://ai.google.dev/gemma/docs/base) **Terms of Use**: [Terms][terms] **Authors**: miner41612 ## Model Information 입력 및 출력에 대한 요약 설명과 간략한 정의입니다. ### Description Google의 Gemma 2 2b 모델을 금융 도메인 데이터셋을 정재한 데이터셋을 Continual Learning을 하여 학습 한 모델에 금융 도메인 Insturction 데이터 셋으로 학습 시킨 모델입니다. ### Usage Below we share some code snippets on how to get quickly started with running the model. First, install the Transformers library with: ```sh pip install -U transformers ``` Then, copy the snippet from the section that is relevant for your usecase. #### Running with the `pipeline` API ```python import torch from transformers import pipeline pipe = pipeline( "text-generation", model="miner41612/gemma-2-2b-finance-it-v1", model_kwargs={"torch_dtype": torch.bfloat16}, device="cuda", # replace with "mps" to run on a Mac device ) messages = [ {"role": "user", "content": "원가상환제도란?"}, ] outputs = pipe(messages, max_new_tokens=256) assistant_response = outputs[0]["generated_text"][-1]["content"].strip() print(assistant_response) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", device_map="auto", ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` You can ensure the correct chat template is applied by using `tokenizer.apply_chat_template` as follows: ```python messages = [ {"role": "user", "content": "원가상환제도란?"}, ] input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True).to("cuda") outputs = model.generate(**input_ids, max_new_tokens=256) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` <details> <summary> Using 8-bit precision (int8) </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> <details> <summary> Using 4-bit precision </summary> ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = AutoModelForCausalLM.from_pretrained( "miner41612/gemma-2-2b-finance-it-v1", quantization_config=quantization_config, ) input_text = "원가상환제도란?" input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids, max_new_tokens=32) print(tokenizer.decode(outputs[0])) ``` </details> #### Advanced Usage <details> <summary> Torch compile </summary> [Torch compile](https://pytorch.org/tutorials/intermediate/torch_compile_tutorial.html) is a method for speeding-up the inference of PyTorch modules. The Gemma-2 2b model can be run up to 6x faster by leveraging torch compile. Note that two warm-up steps are required before the full inference speed is realised: ```python import os os.environ["TOKENIZERS_PARALLELISM"] = "false" from transformers import AutoTokenizer, Gemma2ForCausalLM from transformers.cache_utils import HybridCache import torch torch.set_float32_matmul_precision("high") # load the model + tokenizer tokenizer = AutoTokenizer.from_pretrained("miner41612/gemma-2-2b-finance-it-v1") model = Gemma2ForCausalLM.from_pretrained("miner41612/gemma-2-2b-finance-it-v1", torch_dtype=torch.bfloat16) model.to("cuda") # apply the torch compile transformation model.forward = torch.compile(model.forward, mode="reduce-overhead", fullgraph=True) # pre-process inputs input_text = "원가상환제도란? " model_inputs = tokenizer(input_text, return_tensors="pt").to("cuda") prompt_length = model_inputs.input_ids.shape[1] # set-up k/v cache past_key_values = HybridCache( config=model.config, max_batch_size=1, max_cache_len=model.config.max_position_embeddings, device=model.device, dtype=model.dtype ) # enable passing kv cache to generate model._supports_cache_class = True model.generation_config.cache_implementation = None # two warm-up steps for idx in range(2): outputs = model.generate(**model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) past_key_values.reset() # fast run outputs = model.generate(**model_inputs, past_key_values=past_key_values, do_sample=True, temperature=1.0, max_new_tokens=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` For more details, refer to the [Transformers documentation](https://huggingface.co/docs/transformers/main/en/llm_optims?static-kv=basic+usage%3A+generation_config). </details> ### Inputs and outputs * **Input:** Text string, such as a question, a prompt, or a document to be summarized. * **Output:** Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ### Citation ```none @article{gemma_2024, title={Gemma}, url={https://www.kaggle.com/m/3301}, DOI={10.34740/KAGGLE/M/3301}, publisher={Kaggle}, author={Gemma Team}, year={2024} } ``` ## Model Data Data used for model training and how the data was processed. ## Ethics and Safety Ethics and safety evaluation approach and results. ## Dangerous Capability Evaluations ### Evaluation Approach We evaluated a range of dangerous capabilities: - **Offensive cybersecurity:** To assess the model's potential for misuse in cybersecurity contexts, we utilized both publicly available Capture-the-Flag (CTF) platforms like InterCode-CTF and Hack the Box, as well as internally developed CTF challenges. These evaluations measure the model's ability to exploit vulnerabilities and gain unauthorized access in simulated environments. - **Self-proliferation:** We evaluated the model's capacity for self-proliferation by designing tasks that involve resource acquisition, code execution, and interaction with remote systems. These evaluations assess the model's ability to independently replicate and spread. - **Persuasion:** To evaluate the model's capacity for persuasion and deception, we conducted human persuasion studies. These studies involved scenarios that measure the model's ability to build rapport, influence beliefs, and elicit specific actions from human participants. ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit][rai-toolkit]. * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy][prohibited-use]. * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques.

{}

task

[ "SUMMARIZATION" ]

44,595

9unu/formal_speech_translation

9unu

text2text-generation

[ "transformers", "safetensors", "t5", "text2text-generation", "license:mit", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2024-05-13T10:07:45Z

2024-05-27T10:12:59+00:00

4

0

--- license: mit --- # formal_speech_translation 카카오톡 AI 말투 변환기의 역공학 엔지니어링을 통해 개발한 **대화체에 견고한 상냥체 변환기**입니다. [t5모델](https://huggingface.co/google-t5)을 한국어 전용 데이터로 학습한 [pko-t5-base](https://huggingface.co/paust/pko-t5-base) 모델을 활용하여 학습을 진행했습니다. [dev_repository](https://github.com/9unu/Kakao_Reverse-Engineering) ## Usage transformers의 pipe, from_pretrained 등의 api를 활용하여 접근 가능합니다. ## Example ```python from transformers import T5TokenizerFast, T5ForConditionalGeneration, pipeline # 모델 경로 및 cpu, gpu 지정 cache_dir = "./hugging_face" gentle_model_path='9unu/formal_speech_translation' gentle_model = T5ForConditionalGeneration.from_pretrained(formal_model_path, cache_dir=cache_dir) device = torch.device("cuda" if torch.cuda.is_available() else "cpu") ``` ```python # transformers 파이프라인 생성 gentle_pipeline = pipeline(model = gentle_model, tokenizer = tokenizer, device = device, max_length=60) # text 말투 변환 text = "밥 먹는 중이야" num_return_sequences = 1 max_length = 60 out = gentle_pipeline(text, num_return_sequences = num_return_sequences, max_length=max_length) print([x['generated_text'] for x in out]) ``` ## License 본 모델은 MIT license 하에 공개되어 있습니다.

null

Non_BioNLP

# formal_speech_translation 카카오톡 AI 말투 변환기의 역공학 엔지니어링을 통해 개발한 **대화체에 견고한 상냥체 변환기**입니다. [t5모델](https://huggingface.co/google-t5)을 한국어 전용 데이터로 학습한 [pko-t5-base](https://huggingface.co/paust/pko-t5-base) 모델을 활용하여 학습을 진행했습니다. [dev_repository](https://github.com/9unu/Kakao_Reverse-Engineering) ## Usage transformers의 pipe, from_pretrained 등의 api를 활용하여 접근 가능합니다. ## Example ```python from transformers import T5TokenizerFast, T5ForConditionalGeneration, pipeline # 모델 경로 및 cpu, gpu 지정 cache_dir = "./hugging_face" gentle_model_path='9unu/formal_speech_translation' gentle_model = T5ForConditionalGeneration.from_pretrained(formal_model_path, cache_dir=cache_dir) device = torch.device("cuda" if torch.cuda.is_available() else "cpu") ``` ```python # transformers 파이프라인 생성 gentle_pipeline = pipeline(model = gentle_model, tokenizer = tokenizer, device = device, max_length=60) # text 말투 변환 text = "밥 먹는 중이야" num_return_sequences = 1 max_length = 60 out = gentle_pipeline(text, num_return_sequences = num_return_sequences, max_length=max_length) print([x['generated_text'] for x in out]) ``` ## License 본 모델은 MIT license 하에 공개되어 있습니다.

{"license": "mit"}

task

[ "TRANSLATION" ]

44,596

somosnlp/gemma-7b-it-legal-refugiados-es

somosnlp

null

[ "transformers, pe", "safetensors", "gemma", "trl", "sft", "generated_from_trainer", "es", "dataset:somosnlp/instruct-legal-refugiados-es", "arxiv:1910.09700", "base_model:google/gemma-7b", "base_model:finetune:google/gemma-7b", "license:apache-2.0", "region:us" ]

2024-03-18T22:59:55Z

2024-04-24T11:36:00+00:00

16

0

--- base_model: google/gemma-7b datasets: - somosnlp/instruct-legal-refugiados-es language: - es library_name: transformers, pe license: apache-2.0 tags: - trl - sft - generated_from_trainer ---  # Model Card for gemma-7b-it-legal-refugiados-es    Spain is the third country with the highest number of asylum applications, receiving each year approximately more than 100,000 applications, and the third with the lowest number of approvals within the EU. The main objective of this project is to facilitate the tasks of NGOs in this field and other institutions and help them to obtain answers to questions (QA) related to refugee legislation in Spanish. With its refined understanding of the nuances and complexities of this legal field. The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain ## Model Details ### Model Description  The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the dataset [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es). This is the model card of a 🤗 transformers model that has been pushed on the Hub to allow public access. - **Developed by:**  [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim) - **Funded by:** SomosNLP, HuggingFace  - **Model type:** Language model, instruction tuned - **Language(s):** es-ES, es-MX, es-VE  - **License:** apache-2.0  - **Fine-tuned from model:** [google/gemma-7b](https://huggingface.co/google/  - **Dataset used:** [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es)  ### Model Sources - **Repository:** Notebook in [This repo](https://huggingface.co/somosnlp/gemma-7b-it-legal-refugee-v0.1.1)  - **Demo:** [Demo Space](https://huggingface.co/spaces/somosnlp/QA-legal-refugiados)  - **Video presentation:** [Youtube Video](https://www.youtube.com/watch?v=1OqHDE5LKMI&list=PLTA-KAy8nxaASMwEUWkkTfMaDxWBxn-8J&index=3)  ### Model Family  This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b). ## Uses  ### Direct Use  The primary objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. ### Downstream Use  Intented to be use in question-answering with a context and text generation. ### Out-of-Scope Use  Misuse includes any application that promotes unethical practices, misinterprets refugee law, or uses the model for malicious purposes. The model is not designed to replace professional legal advice. ## Bias, Risks, and Limitations  The model, while powerful, has limitations inherent to AI, including biases present in the training data. It may not cover all nuances of refugee regulations or adapt to changes in law without updates. ### Recommendations   ## How to Get Started with the Model Use the code below to get started with the model. ```python import torch from transformers import ( AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, pipeline ) model_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer = AutoTokenizer.from_pretrained(tokenizer_id) # Cargamos el modelo en 4 bits para agilizar la inferencia quantization_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16, bnb_4bit_use_double_quant=True, ) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, device_map="auto", quantization_config=quantization_config, ) # Generamos el pipeline de generación de texto pipe = pipeline("text-generation", model=model, tokenizer=tokenizer) # Definimos el eos token para el modelo eos_token = tokenizer("<|im_end|>",add_special_tokens=False)["input_ids"][0] def generate_inference(instruction, input, temperature): prompt = pipe.tokenizer.apply_chat_template([{"role": "user", "content": f"{instruction}/n{input}"}], tokenize=False, add_generation_prompt=True) outputs = pipe(prompt, max_new_tokens=256, do_sample=True, num_beams=1, temperature=float(temperature), top_k=50, top_p=0.95, max_time= 300, eos_token_id=eos_token) return outputs[0]['generated_text'][len(prompt):].strip() instruction = "¿Podrías explicarme brevemente los hechos que originan el procedimiento y las posibles calificaciones, así como las sanciones correspondientes, según lo expuesto en el contexto?" input = "b) Hechos que motivan la incoación del procedimiento sucintamente expuestos, su posible calificación y las sanciones que pudieran corresponder, sin perjuicio de lo que resulte de la instrucción. c) Instructor y, en su caso, secretario del procedimiento, con expresa indicación del régimen de recusación de éstos. d) Órgano competente para la resolución del expediente y norma que le atribuye tal competencia. e) Indicación de la posibilidad de que el presunto responsable pueda reconocer voluntariamente su responsabilidad. f) Medidas de carácter provisional que se hayan acordado por el órgano competente para iniciar el procedimiento sancionador, sin perjuicio de las que se puedan adoptar durante éste de conformidad con los artículos 55 y 61 de la Ley Orgánica 4/2000, de 11 de enero. g) Indicación del derecho a formular alegaciones y a la audiencia en el procedimiento y de los plazos para su ejercicio. 2. El acuerdo de iniciación se comunicará al instructor con traslado de cuantas actuaciones existan al respecto y se notificará a los interesados, entendiéndose en todo caso por tal al expedientado. En la notificación se advertirá a los interesados que, de no efectuar alegaciones sobre el contenido de la iniciación del procedimiento en el plazo previsto en el artículo siguiente, no realizarse propuesta de prueba o no ser admitidas, por improcedentes o innecesarias, las pruebas propuestas, la iniciación podrá ser considerada propuesta de resolución cuando contenga un pronunciamiento preciso acerca de la responsabilidad imputada, con los efectos previstos en los artículos 229 y 230." response = test_inference(instruction, input, 0.3) print(f"Response:\n{response}") ``` ## Training Details ### Training Data  The dataset used was [instruct-legal-refugiados-es](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es) but we adapted the dataset to a ChatML format, described in the next section. ### Training Procedure   The training was done using RTX 4090 from Vast.ai with PeRF and Lora #### Preprocessing We wanted to make a conversation model so we investigated the base model prompt in order to make conversational base on [chatml format](https://github.com/MicrosoftDocs/azure-docs/blob/main/articles/ai-services/openai/includes/chat-markup-language.md#working-with-chat-markup-language-chatml) we identified the special tokens so the model could understand the different roles in the conversation Example ``` <bos><|im_start|>system You are Gemma.<|im_end|> <|im_start|>user Hello, how are you?<|im_end|> <|im_start|>assistant I'm doing great. How can I help you today?<|im_end|>\n<eos> ``` So we used [Phil Schmid's gemma chatml tokenizer](https://huggingface.co/philschmid/gemma-tokenizer-chatml) to adapt our dataset for training #### Training Hyperparameters  The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 66 - gradient_accumulation_steps: 2 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 3 - **Training regime:**  ## Evaluation  ### Testing Data, Factors & Metrics #### Testing Data  [More Information Needed] #### Factors  [More Information Needed] #### Metrics  [More Information Needed] ### Results  [More Information Needed] ## Environmental Impact   Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - **Hardware Type**: 1 X RTX4090 - **Hours used**: 4 - **Cloud Provider**: Vast.ai - **Compute Region**: West Europe - **Carbon Emitted**: 350W x 4h = 1.4 kWh x 0.57 kg eq. CO2/kWh = 0.8 kg eq. CO2 ## Technical Specifications  ### Model Architecture and Objective The base model is [google/gemma-7b](https://huggingface.co/google/gemma-7b) finetuned in 4-bit. ### Compute Infrastructure #### Hardware  1 x RTX4090 GPU by Vast.ai. #### Software  Libraries: - transformers - bitsandbytes - accelerate - xformers - trl - peft - wandb ## License  This model is under the license of the Gemma models by Google. Link to consent: https://www.kaggle.com/models/google/gemma/license/consent ## Citation **BibTeX:** [More Information Needed]  ``` @software{somosnlp2024asistenciarefugiados, author = {Alvaro Hidalgo, Eduardo Muñoz, Teresa Martín}, title = {gemma-7b-it-legal-refugiados-es}, month = April, year = 2024, url = {somosnlp/gemma-7b-it-legal-refugee-v0.1.1} } ``` ## More Information  This project was developed during the [Hackathon #Somos600M](https://somosnlp.org/hackathon) organized by SomosNLP. The model was trained using GPUs sponsored by HuggingFace. **Team:** [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim)  ## Contact [optional]

null

Non_BioNLP

# Model Card for gemma-7b-it-legal-refugiados-es    Spain is the third country with the highest number of asylum applications, receiving each year approximately more than 100,000 applications, and the third with the lowest number of approvals within the EU. The main objective of this project is to facilitate the tasks of NGOs in this field and other institutions and help them to obtain answers to questions (QA) related to refugee legislation in Spanish. With its refined understanding of the nuances and complexities of this legal field. The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain ## Model Details ### Model Description  The objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the dataset [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es). This is the model card of a 🤗 transformers model that has been pushed on the Hub to allow public access. - **Developed by:**  [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim) - **Funded by:** SomosNLP, HuggingFace  - **Model type:** Language model, instruction tuned - **Language(s):** es-ES, es-MX, es-VE  - **License:** apache-2.0  - **Fine-tuned from model:** [google/gemma-7b](https://huggingface.co/google/  - **Dataset used:** [AsistenciaRefugiados](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es)  ### Model Sources - **Repository:** Notebook in [This repo](https://huggingface.co/somosnlp/gemma-7b-it-legal-refugee-v0.1.1)  - **Demo:** [Demo Space](https://huggingface.co/spaces/somosnlp/QA-legal-refugiados)  - **Video presentation:** [Youtube Video](https://www.youtube.com/watch?v=1OqHDE5LKMI&list=PLTA-KAy8nxaASMwEUWkkTfMaDxWBxn-8J&index=3)  ### Model Family  This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b). ## Uses  ### Direct Use  The primary objective of this model is to facilitate question answering (QA) tasks pertaining to Spanish refugee legislation. With its refined understanding of the nuances and intricacies of this legal domain. ### Downstream Use  Intented to be use in question-answering with a context and text generation. ### Out-of-Scope Use  Misuse includes any application that promotes unethical practices, misinterprets refugee law, or uses the model for malicious purposes. The model is not designed to replace professional legal advice. ## Bias, Risks, and Limitations  The model, while powerful, has limitations inherent to AI, including biases present in the training data. It may not cover all nuances of refugee regulations or adapt to changes in law without updates. ### Recommendations   ## How to Get Started with the Model Use the code below to get started with the model. ```python import torch from transformers import ( AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, pipeline ) model_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer_id = "somosnlp/gemma-7b-it-legal-refugiados-es" tokenizer = AutoTokenizer.from_pretrained(tokenizer_id) # Cargamos el modelo en 4 bits para agilizar la inferencia quantization_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16, bnb_4bit_use_double_quant=True, ) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, device_map="auto", quantization_config=quantization_config, ) # Generamos el pipeline de generación de texto pipe = pipeline("text-generation", model=model, tokenizer=tokenizer) # Definimos el eos token para el modelo eos_token = tokenizer("<|im_end|>",add_special_tokens=False)["input_ids"][0] def generate_inference(instruction, input, temperature): prompt = pipe.tokenizer.apply_chat_template([{"role": "user", "content": f"{instruction}/n{input}"}], tokenize=False, add_generation_prompt=True) outputs = pipe(prompt, max_new_tokens=256, do_sample=True, num_beams=1, temperature=float(temperature), top_k=50, top_p=0.95, max_time= 300, eos_token_id=eos_token) return outputs[0]['generated_text'][len(prompt):].strip() instruction = "¿Podrías explicarme brevemente los hechos que originan el procedimiento y las posibles calificaciones, así como las sanciones correspondientes, según lo expuesto en el contexto?" input = "b) Hechos que motivan la incoación del procedimiento sucintamente expuestos, su posible calificación y las sanciones que pudieran corresponder, sin perjuicio de lo que resulte de la instrucción. c) Instructor y, en su caso, secretario del procedimiento, con expresa indicación del régimen de recusación de éstos. d) Órgano competente para la resolución del expediente y norma que le atribuye tal competencia. e) Indicación de la posibilidad de que el presunto responsable pueda reconocer voluntariamente su responsabilidad. f) Medidas de carácter provisional que se hayan acordado por el órgano competente para iniciar el procedimiento sancionador, sin perjuicio de las que se puedan adoptar durante éste de conformidad con los artículos 55 y 61 de la Ley Orgánica 4/2000, de 11 de enero. g) Indicación del derecho a formular alegaciones y a la audiencia en el procedimiento y de los plazos para su ejercicio. 2. El acuerdo de iniciación se comunicará al instructor con traslado de cuantas actuaciones existan al respecto y se notificará a los interesados, entendiéndose en todo caso por tal al expedientado. En la notificación se advertirá a los interesados que, de no efectuar alegaciones sobre el contenido de la iniciación del procedimiento en el plazo previsto en el artículo siguiente, no realizarse propuesta de prueba o no ser admitidas, por improcedentes o innecesarias, las pruebas propuestas, la iniciación podrá ser considerada propuesta de resolución cuando contenga un pronunciamiento preciso acerca de la responsabilidad imputada, con los efectos previstos en los artículos 229 y 230." response = test_inference(instruction, input, 0.3) print(f"Response:\n{response}") ``` ## Training Details ### Training Data  The dataset used was [instruct-legal-refugiados-es](https://huggingface.co/datasets/somosnlp/instruct-legal-refugiados-es) but we adapted the dataset to a ChatML format, described in the next section. ### Training Procedure   The training was done using RTX 4090 from Vast.ai with PeRF and Lora #### Preprocessing We wanted to make a conversation model so we investigated the base model prompt in order to make conversational base on [chatml format](https://github.com/MicrosoftDocs/azure-docs/blob/main/articles/ai-services/openai/includes/chat-markup-language.md#working-with-chat-markup-language-chatml) we identified the special tokens so the model could understand the different roles in the conversation Example ``` <bos><|im_start|>system You are Gemma.<|im_end|> <|im_start|>user Hello, how are you?<|im_end|> <|im_start|>assistant I'm doing great. How can I help you today?<|im_end|>\n<eos> ``` So we used [Phil Schmid's gemma chatml tokenizer](https://huggingface.co/philschmid/gemma-tokenizer-chatml) to adapt our dataset for training #### Training Hyperparameters  The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 66 - gradient_accumulation_steps: 2 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 3 - **Training regime:**  ## Evaluation  ### Testing Data, Factors & Metrics #### Testing Data  [More Information Needed] #### Factors  [More Information Needed] #### Metrics  [More Information Needed] ### Results  [More Information Needed] ## Environmental Impact   Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - **Hardware Type**: 1 X RTX4090 - **Hours used**: 4 - **Cloud Provider**: Vast.ai - **Compute Region**: West Europe - **Carbon Emitted**: 350W x 4h = 1.4 kWh x 0.57 kg eq. CO2/kWh = 0.8 kg eq. CO2 ## Technical Specifications  ### Model Architecture and Objective The base model is [google/gemma-7b](https://huggingface.co/google/gemma-7b) finetuned in 4-bit. ### Compute Infrastructure #### Hardware  1 x RTX4090 GPU by Vast.ai. #### Software  Libraries: - transformers - bitsandbytes - accelerate - xformers - trl - peft - wandb ## License  This model is under the license of the Gemma models by Google. Link to consent: https://www.kaggle.com/models/google/gemma/license/consent ## Citation **BibTeX:** [More Information Needed]  ``` @software{somosnlp2024asistenciarefugiados, author = {Alvaro Hidalgo, Eduardo Muñoz, Teresa Martín}, title = {gemma-7b-it-legal-refugiados-es}, month = April, year = 2024, url = {somosnlp/gemma-7b-it-legal-refugee-v0.1.1} } ``` ## More Information  This project was developed during the [Hackathon #Somos600M](https://somosnlp.org/hackathon) organized by SomosNLP. The model was trained using GPUs sponsored by HuggingFace. **Team:** [Alvaro Hidalgo](https://huggingface.co/hacendado) [Eduardo Muñoz](https://huggingface.co/edumunozsala) [Teresa Martin](https://huggingface.co/narhim)  ## Contact [optional]

{"base_model": "google/gemma-7b", "datasets": ["somosnlp/instruct-legal-refugiados-es"], "language": ["es"], "library_name": "transformers, pe", "license": "apache-2.0", "tags": ["trl", "sft", "generated_from_trainer"]}

task

[ "QUESTION_ANSWERING" ]

44,597

MugheesAwan11/bge-base-securiti-dataset-1-v10

MugheesAwan11

sentence-similarity

[ "sentence-transformers", "safetensors", "bert", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:900", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss", "en", "arxiv:1908.10084", "arxiv:2205.13147", "arxiv:1705.00652", "base_model:BAAI/bge-base-en-v1.5", "base_model:finetune:BAAI/bge-base-en-v1.5", "license:apache-2.0", "model-index", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

2024-06-13T17:25:39Z

2024-06-13T17:25:56+00:00

6

0

--- base_model: BAAI/bge-base-en-v1.5 datasets: [] language: - en library_name: sentence-transformers license: apache-2.0 metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:900 - loss:MatryoshkaLoss - loss:MultipleNegativesRankingLoss widget: - source_sentence: Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about Securiti, our mission and history Partner Program Join our Partner Program Contact Us Contact us to learn more or schedule a demo News Coverage Read about Securiti in the news Press Releases Find our latest press releases Careers Join the sentences: - What is the purpose of tracking changes and transformations of data throughout its lifecycle? - What is the role of ePD in the European privacy regime and its relation to GDPR? - How can data governance be optimized using granular insights? - source_sentence: Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud | Data Mapping | DSR Automation | Assessment Automation | Vendor Assessment | Breach Management | Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data | People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent | Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator , View Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud | Data Mapping | DSR Automation | Assessment Automation | Vendor Assessment | Breach Management | Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data | People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent | Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls sentences: - What is the purpose of Asset and Data Discovery in data governance and security? - Which EU member states have strict cyber laws? - What is the obligation for organizations to provide Data Protection Impact Assessments (DPIAs) under the LGPD? - source_sentence: 'which the data is processed. **Right to Access:** Data subjects have the right to obtain confirmation whether or not the controller holds personal data about them, access their personal data, and obtain descriptions of data recipients. **Right to Rectification** : Under the right to rectification, data subjects can request the correction of their data. **Right to Erasure:** Data subjects have the right to request the erasure and destruction of the data that is no longer needed by the organization. **Right to Object:** The data subject has the right to prevent the data controller from processing personal data if such processing causes or is likely to cause unwarranted damage or distress to the data subject. **Right not to be Subjected to Automated Decision-Making** : The data subject has the right to not be subject to automated decision-making that significantly affects the individual. ## Facts related to Ghana’s Data Protection Act 2012 1 While processing personal data, organizations must comply with eight privacy principles: lawfulness of processing, data quality, security measures, accountability, purpose specification, purpose limitation, openness, and data subject participation. 2 In the event of a security breach, the data controller shall take measures to prevent the breach and notify the Commission and the data subject about the breach as soon as reasonably practicable after the discovery of the breach. 3 The DPA specifies lawful grounds for data processing, including data subject’s consent, the performance of a contract, the interest of data subject and public interest, lawful obligations, and the legitimate interest of the data controller. 4 The DPA requires data controllers to register with the Data Protection Commission (DPC). 5 The DPA provides varying fines and terms of imprisonment according to the severity and sensitivity of the violation, such as any person who sells personal data may get fined up to 2500 penalty units or up to five years imprisonment or both. ### Forrester Names Securiti a Leader in the Privacy Management Wave Q4, 2021 Read the Report ### Securiti named a Leader in the IDC MarketScape for Data Privacy Compliance Software Read the Report At Securiti, our mission is to enable enterprises to safely harness the incredible power of data and the cloud by controlling the complex security, privacy and compliance risks. Copyright (C) 2023 Securiti Sitem' sentences: - What information is required for data subjects regarding data transfers under the GDPR, including personal data categories, data recipients, retention period, and automated decision making? - What privacy principles must organizations follow when processing personal data under Ghana's Data Protection Act 2012? - What is the purpose of Thailand's PDPA? - source_sentence: 'consumer has the right to have his/her personal data stored or processed by the data controller be deleted. ## Portability The consumer has a right to obtain a copy of his/her personal data in a portable, technically feasible and readily usable format that allows the consumer to transmit the data to another controller without hindrance. ## Opt out The consumer has the right to opt out of the processing of the personal data for purposes of targeted advertising, the sale of personal data, or profiling in furtherance of decisions that produce legal or similarly significant effects concerning the consumer. **Time period to fulfill DSR request: ** All data subject rights’ requests (DSR requests) must be fulfilled by the data controller within a 45 day period. **Extension in time period: ** data controllers may seek for an extension of 45 days in fulfilling the request depending on the complexity and number of the consumer''s requests. **Denial of DSR request: ** If a DSR request is to be denied, the data controller must inform the consumer of the reasons within a 45 days period. **Appeal against refusal: ** Consumers have a right to appeal the decision for refusal of grant of the DSR request. The appeal must be decided within 45 days but the time period can be further extended by 60 additional days. **Limitation of DSR requests per year: ** Requests for data portability may be made only twice in a year. **Charges: ** DSR requests must be fulfilled free of charge once in a year. Any subsequent request within a 12 month period can be charged. **Authentication: ** A data controller is not to respond to a consumer request unless it can authenticate the request using reasonably commercial means. A data controller can request additional information from the consumer for the purposes of authenticating the request. ## Who must comply? CPA applies to all data controllers who conduct business in Colorado or produce or deliver commercial products or services that are intentionally targeted to residents of Colorado if they match any one or both of these conditions: If they control or process the personal data of 100,000 consumers or more during a calendar year; or If they derive revenue or receive a discount on the price of goods or services from the sale of personal data and process or control the personal data of 25,000' sentences: - What is the US California CCPA and how does it relate to data privacy regulations? - What does the People Data Graph serve in terms of privacy, security, and governance? - What rights does a consumer have regarding the portability of their personal data? - source_sentence: 'PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law''s provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany''s Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR''s €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country''s data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company''s employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users'' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory' sentences: - What are the benefits of automating compliance with global privacy regulations for data protection and control? - What is required for an official complaint to be filed under Germany's Federal Data Protection Act? - Why is tracking data lineage important for data management and security? model-index: - name: SentenceTransformer based on BAAI/bge-base-en-v1.5 results: - task: type: information-retrieval name: Information Retrieval dataset: name: dim 512 type: dim_512 metrics: - type: cosine_accuracy@1 value: 0.08 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.29 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.48 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.65 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.08 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.09666666666666668 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.09599999999999997 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.06499999999999999 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.08 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.29 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.48 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.65 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.3356834483699582 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.23805952380952378 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.25373588653956675 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 256 type: dim_256 metrics: - type: cosine_accuracy@1 value: 0.09 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.33 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.52 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.68 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.09 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.11 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.10399999999999998 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.06799999999999998 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.09 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.33 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.52 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.68 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.35403179411423247 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.2524960317460317 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.26470102220887337 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 128 type: dim_128 metrics: - type: cosine_accuracy@1 value: 0.09 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.27 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.45 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.65 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.09 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.09 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.09 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.06499999999999999 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.09 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.27 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.45 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.65 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.33203261209382817 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.23417063492063486 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.24858408269645846 name: Cosine Map@100 - task: type: information-retrieval name: Information Retrieval dataset: name: dim 64 type: dim_64 metrics: - type: cosine_accuracy@1 value: 0.06 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.23 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.44 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.57 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.06 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.07666666666666666 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.08799999999999997 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.056999999999999995 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.06 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.23 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.44 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.57 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.28544770610641695 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.19726587301587298 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.21493811628701745 name: Cosine Map@100 --- # SentenceTransformer based on BAAI/bge-base-en-v1.5 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 tokens - **Similarity Function:** Cosine Similarity  - **Language:** en - **License:** apache-2.0 ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("MugheesAwan11/bge-base-securiti-dataset-1-v10") # Run inference sentences = [ "PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law's provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany's Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR's €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country's data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company's employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory", "What is required for an official complaint to be filed under Germany's Federal Data Protection Act?", 'Why is tracking data lineage important for data management and security?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_512` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.08 | | cosine_accuracy@3 | 0.29 | | cosine_accuracy@5 | 0.48 | | cosine_accuracy@10 | 0.65 | | cosine_precision@1 | 0.08 | | cosine_precision@3 | 0.0967 | | cosine_precision@5 | 0.096 | | cosine_precision@10 | 0.065 | | cosine_recall@1 | 0.08 | | cosine_recall@3 | 0.29 | | cosine_recall@5 | 0.48 | | cosine_recall@10 | 0.65 | | cosine_ndcg@10 | 0.3357 | | cosine_mrr@10 | 0.2381 | | **cosine_map@100** | **0.2537** | #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.09 | | cosine_accuracy@3 | 0.33 | | cosine_accuracy@5 | 0.52 | | cosine_accuracy@10 | 0.68 | | cosine_precision@1 | 0.09 | | cosine_precision@3 | 0.11 | | cosine_precision@5 | 0.104 | | cosine_precision@10 | 0.068 | | cosine_recall@1 | 0.09 | | cosine_recall@3 | 0.33 | | cosine_recall@5 | 0.52 | | cosine_recall@10 | 0.68 | | cosine_ndcg@10 | 0.354 | | cosine_mrr@10 | 0.2525 | | **cosine_map@100** | **0.2647** | #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.09 | | cosine_accuracy@3 | 0.27 | | cosine_accuracy@5 | 0.45 | | cosine_accuracy@10 | 0.65 | | cosine_precision@1 | 0.09 | | cosine_precision@3 | 0.09 | | cosine_precision@5 | 0.09 | | cosine_precision@10 | 0.065 | | cosine_recall@1 | 0.09 | | cosine_recall@3 | 0.27 | | cosine_recall@5 | 0.45 | | cosine_recall@10 | 0.65 | | cosine_ndcg@10 | 0.332 | | cosine_mrr@10 | 0.2342 | | **cosine_map@100** | **0.2486** | #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.06 | | cosine_accuracy@3 | 0.23 | | cosine_accuracy@5 | 0.44 | | cosine_accuracy@10 | 0.57 | | cosine_precision@1 | 0.06 | | cosine_precision@3 | 0.0767 | | cosine_precision@5 | 0.088 | | cosine_precision@10 | 0.057 | | cosine_recall@1 | 0.06 | | cosine_recall@3 | 0.23 | | cosine_recall@5 | 0.44 | | cosine_recall@10 | 0.57 | | cosine_ndcg@10 | 0.2854 | | cosine_mrr@10 | 0.1973 | | **cosine_map@100** | **0.2149** |   ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 900 training samples * Columns: <code>positive</code> and <code>anchor</code> * Approximate statistics based on the first 1000 samples: | | positive | anchor | |:--------|:--------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------| | type | string | string | | details | <ul><li>min: 159 tokens</li><li>mean: 445.26 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 22.05 tokens</li><li>max: 82 tokens</li></ul> | * Samples: | positive | anchor | |:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------| | <code>orra The Andorra personal data protection act came into force on May 17, 2022, by the Andorra Data Protection Authority (ADPA). Learn more about Andorra PDPA ### United Kingdom The UK Data Protection Act (DPA) 2018 is the amended version of the Data Protection Act that was passed in 1998. The DPA 2018 implements the GDPR with several additions and restrictions. Learn more about UK DPA ### Botswana The Botswana Data Protection came into effect on October 15, 2021 after the issuance of the Data Protection Act (Commencement Date) Order 2021 by the Minister of Presidential Affairs, Governance and Public Administration. Learn more about Botswana DPA ### Zambia On March 31, 2021, the Zambian parliament formally passed the Data Protection Act No. 3 of 2021 and the Electronic Communications and Transactions Act No. 4 of 2021. Learn more about Zambia DPA ### Jamaica On November 30, 2020, the First Schedule of the Data Protection Act No. 7 of 2020 came into effect following the publication of Supplement No. 160 of Volume CXLIV in the Jamaica Gazette Supplement. Learn more about Jamaica DPA ### Belarus The Law on Personal Data Protection of May 7, 2021, No. 99-Z, entered into effect within Belarus on November 15, 2021. Learn more about Belarus DPA ### Russian Federation The primary Russian law on data protection, Federal Law No. 152-FZ has been in effect since July 2006. Learn more ### Eswatini On March 4, 2022, the Eswatini Communications Commission published the Data Protection Act No. 5 of 2022, simultaneously announcing its immediate enforcement. Learn more ### Oman The Royal Decree 6/2022 promulgating the Personal Data Protection Law (PDPL) was passed on February 9, 2022. Learn more ### Sri Lanka Sri Lanka's parliament formally passed the Personal Data Protection Act (PDPA), No. 9 Of 2022, on March 19, 2022. Learn more ### Kuwait Kuwait's DPPR was formally introduced by the CITRA to ensure the Gulf country's data privacy infrastructure. Learn more ### Brunei Darussalam The draft Personal Data Protection Order is Brunei’s primary data protection law which came into effect in 2022. Learn more ### India India’</code> | <code>What is the name of India's data protection law before May 17, 2022?</code> | | <code>the affected data subjects and regulatory authority about the breach and whether any of their information has been compromised as a result. ### Data Protection Impact Assessment There is no requirement for conducting data protection impact assessment under the PDPA. ### Record of Processing Activities A data controller must keep and maintain a record of any privacy notice, data subject request, or any other information relating to personal data processed by him in the form and manner that may be determined by the regulatory authority. ### Cross Border Data Transfer Requirements The PDPA provides that personal data can be transferred out of Malaysia only when the recipient country is specified as adequate in the Official Gazette. The personal data of data subjects can not be disclosed without the consent of the data subject. The PDPA provides the following exceptions to the cross border data transfer requirements: Where the consent of data subject is obtained for transfer; or Where the transfer is necessary for the performance of contract between the parties; The transfer is for the purpose of any legal proceedings or for the purpose of obtaining legal advice or for establishing, exercising or defending legal rights; The data user has taken all reasonable precautions and exercised all due diligence to ensure that the personal data will not in that place be processed in any manner which, if that place is Malaysia, would be a contravention of this PDPA; The transfer is necessary in order to protect the vital interests of the data subject; or The transfer is necessary as being in the public interest in circumstances as determined by the Minister. ## Data Subject Rights The data subjects or the person whose data is being collected has certain rights under the PDPA. The most prominent rights can be categorized under the following: ## Right to withdraw consent The PDPA, like some of the other landmark data protection laws such as CPRA and GDPR gives data subjects the right to revoke their consent at any time by way of written notice from having their data collected processed. ## Right to access and rectification As per this right, anyone whose data has been collected has the right to request to review their personal data and have it updated. The onus is on the data handlers to respond to such a request as soon as possible while also making it easier for data subjects on how they can request access to their personal data. ## Right to data portability Data subjects have the right to request that their data be stored in a manner where it</code> | <code>What is the requirement for conducting a data protection impact assessment under the PDPA?</code> | | <code>more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about</code> | <code>What is Data Subject Request Automation?</code> | * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 5 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 5 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs | Epoch | Step | Training Loss | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_512_cosine_map@100 | dim_64_cosine_map@100 | |:-------:|:-------:|:-------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:| | 0.3448 | 10 | 7.4297 | - | - | - | - | | 0.6897 | 20 | 5.5127 | - | - | - | - | | 1.0 | 29 | - | 0.2399 | 0.2435 | 0.2579 | 0.1837 | | 1.0345 | 30 | 4.8788 | - | - | - | - | | 1.3793 | 40 | 4.0614 | - | - | - | - | | 1.7241 | 50 | 3.3471 | - | - | - | - | | 2.0 | 58 | - | 0.2373 | 0.2510 | 0.2545 | 0.1964 | | 2.0690 | 60 | 3.104 | - | - | - | - | | 2.4138 | 70 | 2.695 | - | - | - | - | | 2.7586 | 80 | 2.2038 | - | - | - | - | | 3.0 | 87 | - | 0.2416 | 0.2630 | 0.2587 | 0.2121 | | 3.1034 | 90 | 2.2576 | - | - | - | - | | 3.4483 | 100 | 2.1552 | - | - | - | - | | 3.7931 | 110 | 1.8199 | - | - | - | - | | 4.0 | 116 | - | 0.2429 | 0.2613 | 0.2546 | 0.2098 | | 4.1379 | 120 | 1.9192 | - | - | - | - | | 4.4828 | 130 | 1.7221 | - | - | - | - | | 4.8276 | 140 | 1.6878 | - | - | - | - | | **5.0** | **145** | **-** | **0.2486** | **0.2647** | **0.2537** | **0.2149** | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.1.2+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

null

Non_BioNLP

# SentenceTransformer based on BAAI/bge-base-en-v1.5 This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. ## Model Details ### Model Description - **Model Type:** Sentence Transformer - **Base model:** [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5)  - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 tokens - **Similarity Function:** Cosine Similarity  - **Language:** en - **License:** apache-2.0 ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True}) (2): Normalize() ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("MugheesAwan11/bge-base-securiti-dataset-1-v10") # Run inference sentences = [ "PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law's provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany's Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR's €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country's data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company's employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory", "What is required for an official complaint to be filed under Germany's Federal Data Protection Act?", 'Why is tracking data lineage important for data management and security?', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ```    ## Evaluation ### Metrics #### Information Retrieval * Dataset: `dim_512` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.08 | | cosine_accuracy@3 | 0.29 | | cosine_accuracy@5 | 0.48 | | cosine_accuracy@10 | 0.65 | | cosine_precision@1 | 0.08 | | cosine_precision@3 | 0.0967 | | cosine_precision@5 | 0.096 | | cosine_precision@10 | 0.065 | | cosine_recall@1 | 0.08 | | cosine_recall@3 | 0.29 | | cosine_recall@5 | 0.48 | | cosine_recall@10 | 0.65 | | cosine_ndcg@10 | 0.3357 | | cosine_mrr@10 | 0.2381 | | **cosine_map@100** | **0.2537** | #### Information Retrieval * Dataset: `dim_256` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.09 | | cosine_accuracy@3 | 0.33 | | cosine_accuracy@5 | 0.52 | | cosine_accuracy@10 | 0.68 | | cosine_precision@1 | 0.09 | | cosine_precision@3 | 0.11 | | cosine_precision@5 | 0.104 | | cosine_precision@10 | 0.068 | | cosine_recall@1 | 0.09 | | cosine_recall@3 | 0.33 | | cosine_recall@5 | 0.52 | | cosine_recall@10 | 0.68 | | cosine_ndcg@10 | 0.354 | | cosine_mrr@10 | 0.2525 | | **cosine_map@100** | **0.2647** | #### Information Retrieval * Dataset: `dim_128` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.09 | | cosine_accuracy@3 | 0.27 | | cosine_accuracy@5 | 0.45 | | cosine_accuracy@10 | 0.65 | | cosine_precision@1 | 0.09 | | cosine_precision@3 | 0.09 | | cosine_precision@5 | 0.09 | | cosine_precision@10 | 0.065 | | cosine_recall@1 | 0.09 | | cosine_recall@3 | 0.27 | | cosine_recall@5 | 0.45 | | cosine_recall@10 | 0.65 | | cosine_ndcg@10 | 0.332 | | cosine_mrr@10 | 0.2342 | | **cosine_map@100** | **0.2486** | #### Information Retrieval * Dataset: `dim_64` * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.06 | | cosine_accuracy@3 | 0.23 | | cosine_accuracy@5 | 0.44 | | cosine_accuracy@10 | 0.57 | | cosine_precision@1 | 0.06 | | cosine_precision@3 | 0.0767 | | cosine_precision@5 | 0.088 | | cosine_precision@10 | 0.057 | | cosine_recall@1 | 0.06 | | cosine_recall@3 | 0.23 | | cosine_recall@5 | 0.44 | | cosine_recall@10 | 0.57 | | cosine_ndcg@10 | 0.2854 | | cosine_mrr@10 | 0.1973 | | **cosine_map@100** | **0.2149** |   ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 900 training samples * Columns: <code>positive</code> and <code>anchor</code> * Approximate statistics based on the first 1000 samples: | | positive | anchor | |:--------|:--------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------| | type | string | string | | details | <ul><li>min: 159 tokens</li><li>mean: 445.26 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 22.05 tokens</li><li>max: 82 tokens</li></ul> | * Samples: | positive | anchor | |:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------| | <code>orra The Andorra personal data protection act came into force on May 17, 2022, by the Andorra Data Protection Authority (ADPA). Learn more about Andorra PDPA ### United Kingdom The UK Data Protection Act (DPA) 2018 is the amended version of the Data Protection Act that was passed in 1998. The DPA 2018 implements the GDPR with several additions and restrictions. Learn more about UK DPA ### Botswana The Botswana Data Protection came into effect on October 15, 2021 after the issuance of the Data Protection Act (Commencement Date) Order 2021 by the Minister of Presidential Affairs, Governance and Public Administration. Learn more about Botswana DPA ### Zambia On March 31, 2021, the Zambian parliament formally passed the Data Protection Act No. 3 of 2021 and the Electronic Communications and Transactions Act No. 4 of 2021. Learn more about Zambia DPA ### Jamaica On November 30, 2020, the First Schedule of the Data Protection Act No. 7 of 2020 came into effect following the publication of Supplement No. 160 of Volume CXLIV in the Jamaica Gazette Supplement. Learn more about Jamaica DPA ### Belarus The Law on Personal Data Protection of May 7, 2021, No. 99-Z, entered into effect within Belarus on November 15, 2021. Learn more about Belarus DPA ### Russian Federation The primary Russian law on data protection, Federal Law No. 152-FZ has been in effect since July 2006. Learn more ### Eswatini On March 4, 2022, the Eswatini Communications Commission published the Data Protection Act No. 5 of 2022, simultaneously announcing its immediate enforcement. Learn more ### Oman The Royal Decree 6/2022 promulgating the Personal Data Protection Law (PDPL) was passed on February 9, 2022. Learn more ### Sri Lanka Sri Lanka's parliament formally passed the Personal Data Protection Act (PDPA), No. 9 Of 2022, on March 19, 2022. Learn more ### Kuwait Kuwait's DPPR was formally introduced by the CITRA to ensure the Gulf country's data privacy infrastructure. Learn more ### Brunei Darussalam The draft Personal Data Protection Order is Brunei’s primary data protection law which came into effect in 2022. Learn more ### India India’</code> | <code>What is the name of India's data protection law before May 17, 2022?</code> | | <code>the affected data subjects and regulatory authority about the breach and whether any of their information has been compromised as a result. ### Data Protection Impact Assessment There is no requirement for conducting data protection impact assessment under the PDPA. ### Record of Processing Activities A data controller must keep and maintain a record of any privacy notice, data subject request, or any other information relating to personal data processed by him in the form and manner that may be determined by the regulatory authority. ### Cross Border Data Transfer Requirements The PDPA provides that personal data can be transferred out of Malaysia only when the recipient country is specified as adequate in the Official Gazette. The personal data of data subjects can not be disclosed without the consent of the data subject. The PDPA provides the following exceptions to the cross border data transfer requirements: Where the consent of data subject is obtained for transfer; or Where the transfer is necessary for the performance of contract between the parties; The transfer is for the purpose of any legal proceedings or for the purpose of obtaining legal advice or for establishing, exercising or defending legal rights; The data user has taken all reasonable precautions and exercised all due diligence to ensure that the personal data will not in that place be processed in any manner which, if that place is Malaysia, would be a contravention of this PDPA; The transfer is necessary in order to protect the vital interests of the data subject; or The transfer is necessary as being in the public interest in circumstances as determined by the Minister. ## Data Subject Rights The data subjects or the person whose data is being collected has certain rights under the PDPA. The most prominent rights can be categorized under the following: ## Right to withdraw consent The PDPA, like some of the other landmark data protection laws such as CPRA and GDPR gives data subjects the right to revoke their consent at any time by way of written notice from having their data collected processed. ## Right to access and rectification As per this right, anyone whose data has been collected has the right to request to review their personal data and have it updated. The onus is on the data handlers to respond to such a request as soon as possible while also making it easier for data subjects on how they can request access to their personal data. ## Right to data portability Data subjects have the right to request that their data be stored in a manner where it</code> | <code>What is the requirement for conducting a data protection impact assessment under the PDPA?</code> | | <code>more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about</code> | <code>What is Data Subject Request Automation?</code> | * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: epoch - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 5 - `lr_scheduler_type`: cosine - `warmup_ratio`: 0.1 - `bf16`: True - `tf32`: True - `load_best_model_at_end`: True - `optim`: adamw_torch_fused - `batch_sampler`: no_duplicates #### All Hyperparameters <details><summary>Click to expand</summary> - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: epoch - `prediction_loss_only`: True - `per_device_train_batch_size`: 32 - `per_device_eval_batch_size`: 16 - `per_gpu_train_batch_size`: None - `per_gpu_eval_batch_size`: None - `gradient_accumulation_steps`: 1 - `eval_accumulation_steps`: None - `learning_rate`: 2e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1.0 - `num_train_epochs`: 5 - `max_steps`: -1 - `lr_scheduler_type`: cosine - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `warmup_steps`: 0 - `log_level`: passive - `log_level_replica`: warning - `log_on_each_node`: True - `logging_nan_inf_filter`: True - `save_safetensors`: True - `save_on_each_node`: False - `save_only_model`: False - `restore_callback_states_from_checkpoint`: False - `no_cuda`: False - `use_cpu`: False - `use_mps_device`: False - `seed`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: True - `fp16`: False - `fp16_opt_level`: O1 - `half_precision_backend`: auto - `bf16_full_eval`: False - `fp16_full_eval`: False - `tf32`: True - `local_rank`: 0 - `ddp_backend`: None - `tpu_num_cores`: None - `tpu_metrics_debug`: False - `debug`: [] - `dataloader_drop_last`: False - `dataloader_num_workers`: 0 - `dataloader_prefetch_factor`: None - `past_index`: -1 - `disable_tqdm`: False - `remove_unused_columns`: True - `label_names`: None - `load_best_model_at_end`: True - `ignore_data_skip`: False - `fsdp`: [] - `fsdp_min_num_params`: 0 - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False} - `fsdp_transformer_layer_cls_to_wrap`: None - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch_fused - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `ddp_find_unused_parameters`: None - `ddp_bucket_cap_mb`: None - `ddp_broadcast_buffers`: False - `dataloader_pin_memory`: True - `dataloader_persistent_workers`: False - `skip_memory_metrics`: True - `use_legacy_prediction_loop`: False - `push_to_hub`: False - `resume_from_checkpoint`: None - `hub_model_id`: None - `hub_strategy`: every_save - `hub_private_repo`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `eval_do_concat_batches`: True - `fp16_backend`: auto - `push_to_hub_model_id`: None - `push_to_hub_organization`: None - `mp_parameters`: - `auto_find_batch_size`: False - `full_determinism`: False - `torchdynamo`: None - `ray_scope`: last - `ddp_timeout`: 1800 - `torch_compile`: False - `torch_compile_backend`: None - `torch_compile_mode`: None - `dispatch_batches`: None - `split_batches`: None - `include_tokens_per_second`: False - `include_num_input_tokens_seen`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional </details> ### Training Logs | Epoch | Step | Training Loss | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_512_cosine_map@100 | dim_64_cosine_map@100 | |:-------:|:-------:|:-------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:| | 0.3448 | 10 | 7.4297 | - | - | - | - | | 0.6897 | 20 | 5.5127 | - | - | - | - | | 1.0 | 29 | - | 0.2399 | 0.2435 | 0.2579 | 0.1837 | | 1.0345 | 30 | 4.8788 | - | - | - | - | | 1.3793 | 40 | 4.0614 | - | - | - | - | | 1.7241 | 50 | 3.3471 | - | - | - | - | | 2.0 | 58 | - | 0.2373 | 0.2510 | 0.2545 | 0.1964 | | 2.0690 | 60 | 3.104 | - | - | - | - | | 2.4138 | 70 | 2.695 | - | - | - | - | | 2.7586 | 80 | 2.2038 | - | - | - | - | | 3.0 | 87 | - | 0.2416 | 0.2630 | 0.2587 | 0.2121 | | 3.1034 | 90 | 2.2576 | - | - | - | - | | 3.4483 | 100 | 2.1552 | - | - | - | - | | 3.7931 | 110 | 1.8199 | - | - | - | - | | 4.0 | 116 | - | 0.2429 | 0.2613 | 0.2546 | 0.2098 | | 4.1379 | 120 | 1.9192 | - | - | - | - | | 4.4828 | 130 | 1.7221 | - | - | - | - | | 4.8276 | 140 | 1.6878 | - | - | - | - | | **5.0** | **145** | **-** | **0.2486** | **0.2647** | **0.2537** | **0.2149** | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.14 - Sentence Transformers: 3.0.1 - Transformers: 4.41.2 - PyTorch: 2.1.2+cu121 - Accelerate: 0.31.0 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### MultipleNegativesRankingLoss ```bibtex @misc{henderson2017efficient, title={Efficient Natural Language Response Suggestion for Smart Reply}, author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil}, year={2017}, eprint={1705.00652}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

{"base_model": "BAAI/bge-base-en-v1.5", "datasets": [], "language": ["en"], "library_name": "sentence-transformers", "license": "apache-2.0", "metrics": ["cosine_accuracy@1", "cosine_accuracy@3", "cosine_accuracy@5", "cosine_accuracy@10", "cosine_precision@1", "cosine_precision@3", "cosine_precision@5", "cosine_precision@10", "cosine_recall@1", "cosine_recall@3", "cosine_recall@5", "cosine_recall@10", "cosine_ndcg@10", "cosine_mrr@10", "cosine_map@100"], "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "sentence-similarity", "feature-extraction", "generated_from_trainer", "dataset_size:900", "loss:MatryoshkaLoss", "loss:MultipleNegativesRankingLoss"], "widget": [{"source_sentence": "Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator View Solutions Technologies Covering you everywhere with 1000+ integrations across data systems. Snowflake View AWS View Microsoft 365 View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \\+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product brochures, white papers, infographics, analyst reports and more. Knowledge Center Learn about the data privacy, security and governance landscape. Securiti Education Courses and Certifications for data privacy, security and governance professionals. Company About Us Learn all about Securiti, our mission and history Partner Program Join our Partner Program Contact Us Contact us to learn more or schedule a demo News Coverage Read about Securiti in the news Press Releases Find our latest press releases Careers Join the", "sentences": ["What is the purpose of tracking changes and transformations of data throughout its lifecycle?", "What is the role of ePD in the European privacy regime and its relation to GDPR?", "How can data governance be optimized using granular insights?"]}, {"source_sentence": "Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud | Data Mapping | DSR Automation | Assessment Automation | Vendor Assessment | Breach Management | Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data | People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent | Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls Orchestrator , View Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud | Data Mapping | DSR Automation | Assessment Automation | Vendor Assessment | Breach Management | Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data | People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-time streaming platforms Learn more Data Consent Automation First Party Consent | Third Party & Cookie Consent Learn more Data Security Posture Management Secure sensitive data in hybrid multicloud and SaaS environments Learn more Data Breach Impact Analysis & Response Analyze impact of a data breach and coordinate response per global regulatory obligations Learn more Data Catalog Automatically catalog datasets and enable users to find, understand, trust and access data Learn more Data Lineage Track changes and transformations of data throughout its lifecycle Data Controls Orchestrator View Data Command Center View Sensitive Data Intelligence View Asset Discovery Data Discovery & Classification Sensitive Data Catalog People Data Graph Learn more Privacy Automate compliance with global privacy regulations Data Mapping Automation View Data Subject Request Automation View People Data Graph View Assessment Automation View Cookie Consent View Universal Consent View Vendor Risk Assessment View Breach Management View Privacy Policy Management View Privacy Center View Learn more Security Identify data risk and enable protection & control Data Security Posture Management View Data Access Intelligence & Governance View Data Risk Management View Data Breach Analysis View Learn more Governance Optimize Data Governance with granular insights into your data Data Catalog View Data Lineage View Data Quality View Data Controls", "sentences": ["What is the purpose of Asset and Data Discovery in data governance and security?", "Which EU member states have strict cyber laws?", "What is the obligation for organizations to provide Data Protection Impact Assessments (DPIAs) under the LGPD?"]}, {"source_sentence": "which the data is processed. **Right to Access:** Data subjects have the right to obtain confirmation whether or not the controller holds personal data about them, access their personal data, and obtain descriptions of data recipients. **Right to Rectification** : Under the right to rectification, data subjects can request the correction of their data. **Right to Erasure:** Data subjects have the right to request the erasure and destruction of the data that is no longer needed by the organization. **Right to Object:** The data subject has the right to prevent the data controller from processing personal data if such processing causes or is likely to cause unwarranted damage or distress to the data subject. **Right not to be Subjected to Automated Decision-Making** : The data subject has the right to not be subject to automated decision-making that significantly affects the individual. ## Facts related to Ghana’s Data Protection Act 2012 1 While processing personal data, organizations must comply with eight privacy principles: lawfulness of processing, data quality, security measures, accountability, purpose specification, purpose limitation, openness, and data subject participation. 2 In the event of a security breach, the data controller shall take measures to prevent the breach and notify the Commission and the data subject about the breach as soon as reasonably practicable after the discovery of the breach. 3 The DPA specifies lawful grounds for data processing, including data subject’s consent, the performance of a contract, the interest of data subject and public interest, lawful obligations, and the legitimate interest of the data controller. 4 The DPA requires data controllers to register with the Data Protection Commission (DPC). 5 The DPA provides varying fines and terms of imprisonment according to the severity and sensitivity of the violation, such as any person who sells personal data may get fined up to 2500 penalty units or up to five years imprisonment or both. ### Forrester Names Securiti a Leader in the Privacy Management Wave Q4, 2021 Read the Report ### Securiti named a Leader in the IDC MarketScape for Data Privacy Compliance Software Read the Report At Securiti, our mission is to enable enterprises to safely harness the incredible power of data and the cloud by controlling the complex security, privacy and compliance risks. Copyright (C) 2023 Securiti Sitem", "sentences": ["What information is required for data subjects regarding data transfers under the GDPR, including personal data categories, data recipients, retention period, and automated decision making?", "What privacy principles must organizations follow when processing personal data under Ghana's Data Protection Act 2012?", "What is the purpose of Thailand's PDPA?"]}, {"source_sentence": "consumer has the right to have his/her personal data stored or processed by the data controller be deleted. ## Portability The consumer has a right to obtain a copy of his/her personal data in a portable, technically feasible and readily usable format that allows the consumer to transmit the data to another controller without hindrance. ## Opt out The consumer has the right to opt out of the processing of the personal data for purposes of targeted advertising, the sale of personal data, or profiling in furtherance of decisions that produce legal or similarly significant effects concerning the consumer. **Time period to fulfill DSR request: ** All data subject rights’ requests (DSR requests) must be fulfilled by the data controller within a 45 day period. **Extension in time period: ** data controllers may seek for an extension of 45 days in fulfilling the request depending on the complexity and number of the consumer's requests. **Denial of DSR request: ** If a DSR request is to be denied, the data controller must inform the consumer of the reasons within a 45 days period. **Appeal against refusal: ** Consumers have a right to appeal the decision for refusal of grant of the DSR request. The appeal must be decided within 45 days but the time period can be further extended by 60 additional days. **Limitation of DSR requests per year: ** Requests for data portability may be made only twice in a year. **Charges: ** DSR requests must be fulfilled free of charge once in a year. Any subsequent request within a 12 month period can be charged. **Authentication: ** A data controller is not to respond to a consumer request unless it can authenticate the request using reasonably commercial means. A data controller can request additional information from the consumer for the purposes of authenticating the request. ## Who must comply? CPA applies to all data controllers who conduct business in Colorado or produce or deliver commercial products or services that are intentionally targeted to residents of Colorado if they match any one or both of these conditions: If they control or process the personal data of 100,000 consumers or more during a calendar year; or If they derive revenue or receive a discount on the price of goods or services from the sale of personal data and process or control the personal data of 25,000", "sentences": ["What is the US California CCPA and how does it relate to data privacy regulations?", "What does the People Data Graph serve in terms of privacy, security, and governance?", "What rights does a consumer have regarding the portability of their personal data?"]}, {"source_sentence": "PR and Federal Data Protection Act within Germany; To promote awareness within the public related to the risks, rules, safeguards, and rights concerning the processing of personal data; To handle all complaints raised by data subjects related to data processing in addition to carrying out investigations to find out if any data handler has breached any provisions of the Act; ## Penalties for Non compliance The GDPR already laid down some stringent penalties for companies that would be found in breach of the law's provisions. More importantly, as opposed to other data protection laws such as the CCPA and CPRA, non-compliance with the law also meant penalties. Germany's Federal Data Protection Act has a slightly more lenient take in this regard. Suppose a data handler is found to have fraudulently collected data, processed, shared, or sold data without proper consent from the data subjects, not responded or responded with delay to a data subject request, or failed to inform the data subject of a breach properly. In that case, it can be fined up to €50,000. This is in addition to the GDPR's €20 million or 4% of the total worldwide annual turnover of the preceding financial year, whichever is higher, that any organisation found in breach of the law is subject to. However, for this fine to be applied, either the data subject, the Federal Commissioner, or the regulatory authority must file an official complaint. ## How an Organization Can Operationalize the Law Data handlers processing data inside Germany can remain compliant with the country's data protection law if they fulfill the following conditions: Have a comprehensive privacy policy that educates all users of their rights and how to contact the relevant personnel within the organisation in case of a query Hire a competent Data Protection Officer that understands the GDPR and Federal Data Protection Act thoroughly and can lead compliance efforts within your organisation Ensure all the company's employees and staff are acutely aware of their responsibilities under the law Conduct regular data protection impact assessments as well as data mapping exercises to ensure maximum efficiency in your compliance efforts Notify the relevant authorities of a data breach as soon as possible ## How can Securiti Help Data privacy and compliance have become incredibly vital in earning users' trust globally. Most users now expect most businesses to take all the relevant measures to ensure the data they collect is properly stored, protected, and maintained. Data protection laws have made such efforts legally mandatory", "sentences": ["What are the benefits of automating compliance with global privacy regulations for data protection and control?", "What is required for an official complaint to be filed under Germany's Federal Data Protection Act?", "Why is tracking data lineage important for data management and security?"]}], "model-index": [{"name": "SentenceTransformer based on BAAI/bge-base-en-v1.5", "results": [{"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 512", "type": "dim_512"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.08, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.29, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.48, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.65, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.08, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.09666666666666668, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.09599999999999997, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.06499999999999999, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.08, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.29, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.48, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.65, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.3356834483699582, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.23805952380952378, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.25373588653956675, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 256", "type": "dim_256"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.09, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.33, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.52, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.68, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.09, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.11, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.10399999999999998, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.06799999999999998, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.09, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.33, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.52, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.68, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.35403179411423247, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.2524960317460317, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.26470102220887337, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 128", "type": "dim_128"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.09, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.27, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.45, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.65, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.09, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.09, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.09, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.06499999999999999, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.09, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.27, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.45, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.65, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.33203261209382817, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.23417063492063486, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.24858408269645846, "name": "Cosine Map@100"}]}, {"task": {"type": "information-retrieval", "name": "Information Retrieval"}, "dataset": {"name": "dim 64", "type": "dim_64"}, "metrics": [{"type": "cosine_accuracy@1", "value": 0.06, "name": "Cosine Accuracy@1"}, {"type": "cosine_accuracy@3", "value": 0.23, "name": "Cosine Accuracy@3"}, {"type": "cosine_accuracy@5", "value": 0.44, "name": "Cosine Accuracy@5"}, {"type": "cosine_accuracy@10", "value": 0.57, "name": "Cosine Accuracy@10"}, {"type": "cosine_precision@1", "value": 0.06, "name": "Cosine Precision@1"}, {"type": "cosine_precision@3", "value": 0.07666666666666666, "name": "Cosine Precision@3"}, {"type": "cosine_precision@5", "value": 0.08799999999999997, "name": "Cosine Precision@5"}, {"type": "cosine_precision@10", "value": 0.056999999999999995, "name": "Cosine Precision@10"}, {"type": "cosine_recall@1", "value": 0.06, "name": "Cosine Recall@1"}, {"type": "cosine_recall@3", "value": 0.23, "name": "Cosine Recall@3"}, {"type": "cosine_recall@5", "value": 0.44, "name": "Cosine Recall@5"}, {"type": "cosine_recall@10", "value": 0.57, "name": "Cosine Recall@10"}, {"type": "cosine_ndcg@10", "value": 0.28544770610641695, "name": "Cosine Ndcg@10"}, {"type": "cosine_mrr@10", "value": 0.19726587301587298, "name": "Cosine Mrr@10"}, {"type": "cosine_map@100", "value": 0.21493811628701745, "name": "Cosine Map@100"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,598

danielpark/ko-llama-2-jindo-13b-instruct

danielpark

text-generation

[ "peft", "pytorch", "llama", "dsdanielpark", "llama2", "instruct", "instruction", "jindo", "korean", "translation", "13b", "text-generation", "en", "ko", "dataset:korean-jindo-dataset.json", "arxiv:2307.09288", "region:us" ]

2023-07-27T07:10:35Z

2023-11-13T08:29:33+00:00

0

2

--- datasets: - korean-jindo-dataset.json language: - en - ko library_name: peft pipeline_tag: text-generation tags: - dsdanielpark - llama2 - instruct - instruction - jindo - korean - translation - 13b --- # Since this model is still under development, I recommend not using it until it reaches the development stage 5. Development Status :: 2 - Pre-Alpha Developed by MinWoo Park, 2023, Seoul, South Korea. [Contact: [email protected]](mailto:[email protected]). [![Hits](https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fhuggingface.co%2Fdanielpark%2Fko-llama-2-jindo-13b-instruct&count_bg=%23000000&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=views&edge_flat=false)](https://hits.seeyoufarm.com) # danielpark/llama2-jindo-13b-instruct model card ## `Jindo` is sLLM for construct datasets for LLM `KOLANI`. > **Warning** The training is still in progress. This model is an LLM in various language domains, including Korean translation and correction. Its main purpose is to create a dataset for training the Korean LLM "KOLANI" (which is still undergoing training). Furthermore, since this model has been developed solely by one individual without any external support, the release and improvement process might be relatively slow. Jindo is implemented as sLLM for lightweight purposes, thus focusing primarily on the 7B version for optimization, alongside the 13B version. Using: [QLoRA](https://github.com/artidoro/qlora) ## Model Details The weights you are currently viewing are preliminary checkpoints, and **the official weights have not been released yet.** * **Developed by**: [Minwoo Park](https://github.com/dsdanielpark) * **Backbone Model**: [LLaMA2](https://huggingface.co/meta-llama/Llama-2-7b) [[Paper](https://huggingface.co/papers/2307.09288)] * **Model Jindo Variations**: jindo-instruct * **jindo-instruct Variations**: 2b / 7b / 13b * [danielpark/ko-llama-2-jindo-2b-instruct]() (from LLaMA1) * [danielpark/ko-llama-2-jindo-7b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) (from LLaMA2) * [danielpark/ko-llama-2-jindo-13b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-13b-instruct) (from LLaMA2) * This model targets specific domains, so the 70b model will not be released. * **Quantinized Weight**: 7b-gptq(4bit-128g) * [ko-llama-2-jindo-7b-instruct-4bit-128g-gptq](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq) * **Library**: [HuggingFace Transformers](https://github.com/huggingface/transformers) * **License**: This model is licensed under the Meta's [LLaMA2 license](https://github.com/facebookresearch/llama/blob/main/LICENSE). We plan to check the dataset's license along with the official release, but our primary goal is to aim for a commercial-use release by default. * **Where to send comments**: Instructions on how to provide feedback or comments on a model can be found by opening an issue in the [Hugging Face community's model repository](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) * **Contact**: For questions and comments about the model, please email to me [[email protected]](mailto:[email protected]) ## Web Demo I implement the web demo using several popular tools that allow us to rapidly create web UIs. | model | web ui | quantinized | | --- | --- | --- | | danielpark/ko-llama-2-jindo-7b-instruct. | using [gradio](https://github.com/dsdanielpark/gradio) on [colab](https://colab.research.google.com/drive/1zwR7rz6Ym53tofCGwZZU8y5K_t1r1qqo#scrollTo=p2xw_g80xMsD) | - | | danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq | using [text-generation-webui](https://github.com/oobabooga/text-generation-webui) on [colab](https://colab.research.google.com/drive/19ihYHsyg_5QFZ_A28uZNR_Z68E_09L4G) | gptq | | danielpark/ko-llama-2-jindo-7b-instruct-ggml | [koboldcpp-v1.38](https://github.com/LostRuins/koboldcpp/releases/tag/v1.38) | ggml | ## Dataset Details ### Used Datasets - korean-jindo-dataset - The dataset has not been released yet > No other data was used except for the dataset mentioned above ### Prompt Template ``` ### System: {System} ### User: {User} ### Assistant: {Assistant} ``` ## Hardware and Software * **Hardware** * Under 10b model: Trained using the free T4 GPU resource. * Over 10b model: Utilized a Single A100 on Google Colab. * **Training Factors**: [HuggingFace trainer](https://huggingface.co/docs/transformers/main_classes/trainer) ## Evaluation Results Please refer to the following procedure for the evaluation of the backbone model. Other benchmarking and qualitative evaluations for Korean datasets are still pending. ### Overview - We conducted a performance evaluation based on the tasks being evaluated on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). We evaluated our model on four benchmark datasets, which include `ARC-Challenge`, `HellaSwag`, `MMLU`, and `TruthfulQA`. We used the [lm-evaluation-harness repository](https://github.com/EleutherAI/lm-evaluation-harness), specifically commit [b281b0921b636bc36ad05c0b0b0763bd6dd43463](https://github.com/EleutherAI/lm-evaluation-harness/tree/b281b0921b636bc36ad05c0b0b0763bd6dd43463). ## Usage Please refer to the following information and install the appropriate versions compatible with your enviroments. ``` $ pip install -q accelerate==0.21.0 peft==0.4.0 bitsandbytes==0.40.2 transformers==4.31.0 trl==0.4.7 ``` ```python from transformers import AutoTokenizer import transformers import torch model = "danielpark/ko-llama-2-jindo-13b-instruct" # model = "meta-llama/Llama-2-13b-hf" tokenizer = AutoTokenizer.from_pretrained(model) pipeline = transformers.pipeline( "text-generation", model=model, torch_dtype=torch.float16, device_map="auto", ) sequences = pipeline( 'I liked "Breaking Bad" and "Band of Brothers". Do you have any recommendations of other shows I might like?\n', do_sample=True, top_k=10, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id, max_length=200, ) for seq in sequences: print(f"Result: {seq['generated_text']}") ``` To use the model with the `transformers` library on a machine with GPUs, first make sure you have the `transformers` and `accelerate` libraries installed. ```python %pip install "accelerate>=0.16.0,<1" "transformers[torch]>=4.28.1,<5" "torch>=1.13.1,<2" ``` The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline` found in the model repo [here](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required. Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality. It is also fine to remove it if there is sufficient memory. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-13b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto") ``` You can then use the pipeline to answer instructions: ```python res = generate_text("Explain to me the difference between nuclear fission and fusion.") print(res[0]["generated_text"]) ``` Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer: ```python import torch from instruct_pipeline import InstructionTextGenerationPipeline from transformers import AutoModelForCausalLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", padding_side="left") model = AutoModelForCausalLM.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", device_map="auto", torch_dtype=torch.bfloat16) generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer) ``` ### LangChain Usage To use the pipeline with LangChain, you must set `return_full_text=True`, as LangChain expects the full text to be returned and the default for the pipeline is to only return the new text. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-7b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto", return_full_text=True) ``` You can create a prompt that either has only an instruction or has an instruction with context: ```python from langchain import PromptTemplate, LLMChain from langchain.llms import HuggingFacePipeline # template for an instrution with no input prompt = PromptTemplate( input_variables=["instruction"], template="{instruction}") # template for an instruction with input prompt_with_context = PromptTemplate( input_variables=["instruction", "context"], template="{instruction}\n\nInput:\n{context}") hf_pipeline = HuggingFacePipeline(pipeline=generate_text) llm_chain = LLMChain(llm=hf_pipeline, prompt=prompt) llm_context_chain = LLMChain(llm=hf_pipeline, prompt=prompt_with_context) ``` Example predicting using a simple instruction: ```python print(llm_chain.predict(instruction="Explain to me the difference between nuclear fission and fusion.").lstrip()) ``` Example predicting using an instruction with context: ```python context = """George Washington (February 22, 1732[b] - December 14, 1799) was an American military officer, statesman, and Founding Father who served as the first president of the United States from 1789 to 1797.""" print(llm_context_chain.predict(instruction="When was George Washington president?", context=context).lstrip()) ``` ### Scripts - Prepare evaluation environments: ``` # clone the repository git clone https://github.com/EleutherAI/lm-evaluation-harness.git # check out the specific commit git checkout b281b0921b636bc36ad05c0b0b0763bd6dd43463 # change to the repository directory cd lm-evaluation-harness ``` ## Ethical Issues The Jindo model has not been filtered for harmful, biased, or explicit content. As a result, outputs that do not adhere to ethical norms may be generated during use. Please exercise caution when using the model in research or practical applications. ### Ethical Considerations - There were no ethical issues involved, as we did not include the benchmark test set or the training set in the model's training process. As always, we encourage responsible and ethical use of this model. Please note that while Guanaco strives to provide accurate and helpful responses, it is still crucial to cross-verify the information from reliable sources for knowledge-based queries. ## Contact Me To contact me, you can [mail to me. [email protected]](mailto:[email protected]). ## Model Architecture ```python LlamaForCausalLM( (model): LlamaModel( (embed_tokens): Embedding(32000, 4096, padding_idx=0) (layers): ModuleList( (0-31): 32 x LlamaDecoderLayer( (self_attn): LlamaAttention( (q_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (k_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (v_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (o_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (rotary_emb): LlamaRotaryEmbedding() ) (mlp): LlamaMLP( (gate_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (up_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (down_proj): Linear4bit(in_features=11008, out_features=4096, bias=False) (act_fn): SiLUActivation() ) (input_layernorm): LlamaRMSNorm() (post_attention_layernorm): LlamaRMSNorm() ) ) (norm): LlamaRMSNorm() ) (lm_head): Linear(in_features=4096, out_features=32000, bias=False) ) ``` ## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0 ### License: The licenses of the pretrained models, llama1, and llama2, along with the datasets used, are applicable. For other datasets related to this work, guidance will be provided in the official release. The responsibility for verifying all licenses lies with the user, and the developer assumes no liability, explicit or implied, including legal responsibilities. ### Remark: - The "instruct" in the model name can be omitted, but it is used to differentiate between the backbones of llama2 for chat and general purposes. Additionally, this model is created for a specific purpose, so we plan to fine-tune it with a dataset focused on instructions. ### Naive Cost Estimation Assuming linearity without considering various variables. - 1000 simple prompts - 20 minutes of processing time - Approximately $2 in cost (based on Google Colab's 100 computing units and a single A100 GPU estimated at $10). - Required CPU RAM: 5GB (depending on the training data and dummy size) - Required VRAM: 12-13GB Time ($t$) and cost ($c$) are calculated based on the given information as follows: \[ t(n) = \frac{{20 \text{ minutes}}}{{1000 \text{ prompts}}} \times n \] \[ c(n) = \frac{{2 \text{ dollars}}}{{1000 \text{ prompts}}} \times n \] ```python def calculate_time_cost(num_prompts): total_prompts = 1000 total_time_minutes = 20 total_cost_dollars = 2 time_required = (total_time_minutes / total_prompts) * num_prompts cost_required = (total_cost_dollars / total_prompts) * num_prompts return time_required, cost_required # Example num_prompts = n time, cost = calculate_time_cost(num_prompts) print(f"Time for {num_prompts}: {time} minutes") print(f"Cost for {num_prompts}: {cost} dollar") ``` ### Chinchilla scaling laws The Chinchilla scaling laws focus on optimally scaling training compute but often we also care about inference cost. This tool follows [Harm de Vries’ blog post](https://www.harmdevries.com/post/model-size-vs-compute-overhead/) and visualizes the tradeoff between training comput and inference cost (i.e. model size).

null

Non_BioNLP

# Since this model is still under development, I recommend not using it until it reaches the development stage 5. Development Status :: 2 - Pre-Alpha Developed by MinWoo Park, 2023, Seoul, South Korea. [Contact: [email protected]](mailto:[email protected]). [![Hits](https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fhuggingface.co%2Fdanielpark%2Fko-llama-2-jindo-13b-instruct&count_bg=%23000000&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=views&edge_flat=false)](https://hits.seeyoufarm.com) # danielpark/llama2-jindo-13b-instruct model card ## `Jindo` is sLLM for construct datasets for LLM `KOLANI`. > **Warning** The training is still in progress. This model is an LLM in various language domains, including Korean translation and correction. Its main purpose is to create a dataset for training the Korean LLM "KOLANI" (which is still undergoing training). Furthermore, since this model has been developed solely by one individual without any external support, the release and improvement process might be relatively slow. Jindo is implemented as sLLM for lightweight purposes, thus focusing primarily on the 7B version for optimization, alongside the 13B version. Using: [QLoRA](https://github.com/artidoro/qlora) ## Model Details The weights you are currently viewing are preliminary checkpoints, and **the official weights have not been released yet.** * **Developed by**: [Minwoo Park](https://github.com/dsdanielpark) * **Backbone Model**: [LLaMA2](https://huggingface.co/meta-llama/Llama-2-7b) [[Paper](https://huggingface.co/papers/2307.09288)] * **Model Jindo Variations**: jindo-instruct * **jindo-instruct Variations**: 2b / 7b / 13b * [danielpark/ko-llama-2-jindo-2b-instruct]() (from LLaMA1) * [danielpark/ko-llama-2-jindo-7b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) (from LLaMA2) * [danielpark/ko-llama-2-jindo-13b-instruct](https://huggingface.co/danielpark/ko-llama-2-jindo-13b-instruct) (from LLaMA2) * This model targets specific domains, so the 70b model will not be released. * **Quantinized Weight**: 7b-gptq(4bit-128g) * [ko-llama-2-jindo-7b-instruct-4bit-128g-gptq](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq) * **Library**: [HuggingFace Transformers](https://github.com/huggingface/transformers) * **License**: This model is licensed under the Meta's [LLaMA2 license](https://github.com/facebookresearch/llama/blob/main/LICENSE). We plan to check the dataset's license along with the official release, but our primary goal is to aim for a commercial-use release by default. * **Where to send comments**: Instructions on how to provide feedback or comments on a model can be found by opening an issue in the [Hugging Face community's model repository](https://huggingface.co/danielpark/ko-llama-2-jindo-7b-instruct) * **Contact**: For questions and comments about the model, please email to me [[email protected]](mailto:[email protected]) ## Web Demo I implement the web demo using several popular tools that allow us to rapidly create web UIs. | model | web ui | quantinized | | --- | --- | --- | | danielpark/ko-llama-2-jindo-7b-instruct. | using [gradio](https://github.com/dsdanielpark/gradio) on [colab](https://colab.research.google.com/drive/1zwR7rz6Ym53tofCGwZZU8y5K_t1r1qqo#scrollTo=p2xw_g80xMsD) | - | | danielpark/ko-llama-2-jindo-7b-instruct-4bit-128g-gptq | using [text-generation-webui](https://github.com/oobabooga/text-generation-webui) on [colab](https://colab.research.google.com/drive/19ihYHsyg_5QFZ_A28uZNR_Z68E_09L4G) | gptq | | danielpark/ko-llama-2-jindo-7b-instruct-ggml | [koboldcpp-v1.38](https://github.com/LostRuins/koboldcpp/releases/tag/v1.38) | ggml | ## Dataset Details ### Used Datasets - korean-jindo-dataset - The dataset has not been released yet > No other data was used except for the dataset mentioned above ### Prompt Template ``` ### System: {System} ### User: {User} ### Assistant: {Assistant} ``` ## Hardware and Software * **Hardware** * Under 10b model: Trained using the free T4 GPU resource. * Over 10b model: Utilized a Single A100 on Google Colab. * **Training Factors**: [HuggingFace trainer](https://huggingface.co/docs/transformers/main_classes/trainer) ## Evaluation Results Please refer to the following procedure for the evaluation of the backbone model. Other benchmarking and qualitative evaluations for Korean datasets are still pending. ### Overview - We conducted a performance evaluation based on the tasks being evaluated on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). We evaluated our model on four benchmark datasets, which include `ARC-Challenge`, `HellaSwag`, `MMLU`, and `TruthfulQA`. We used the [lm-evaluation-harness repository](https://github.com/EleutherAI/lm-evaluation-harness), specifically commit [b281b0921b636bc36ad05c0b0b0763bd6dd43463](https://github.com/EleutherAI/lm-evaluation-harness/tree/b281b0921b636bc36ad05c0b0b0763bd6dd43463). ## Usage Please refer to the following information and install the appropriate versions compatible with your enviroments. ``` $ pip install -q accelerate==0.21.0 peft==0.4.0 bitsandbytes==0.40.2 transformers==4.31.0 trl==0.4.7 ``` ```python from transformers import AutoTokenizer import transformers import torch model = "danielpark/ko-llama-2-jindo-13b-instruct" # model = "meta-llama/Llama-2-13b-hf" tokenizer = AutoTokenizer.from_pretrained(model) pipeline = transformers.pipeline( "text-generation", model=model, torch_dtype=torch.float16, device_map="auto", ) sequences = pipeline( 'I liked "Breaking Bad" and "Band of Brothers". Do you have any recommendations of other shows I might like?\n', do_sample=True, top_k=10, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id, max_length=200, ) for seq in sequences: print(f"Result: {seq['generated_text']}") ``` To use the model with the `transformers` library on a machine with GPUs, first make sure you have the `transformers` and `accelerate` libraries installed. ```python %pip install "accelerate>=0.16.0,<1" "transformers[torch]>=4.28.1,<5" "torch>=1.13.1,<2" ``` The instruction following pipeline can be loaded using the `pipeline` function as shown below. This loads a custom `InstructionTextGenerationPipeline` found in the model repo [here](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), which is why `trust_remote_code=True` is required. Including `torch_dtype=torch.bfloat16` is generally recommended if this type is supported in order to reduce memory usage. It does not appear to impact output quality. It is also fine to remove it if there is sufficient memory. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-13b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto") ``` You can then use the pipeline to answer instructions: ```python res = generate_text("Explain to me the difference between nuclear fission and fusion.") print(res[0]["generated_text"]) ``` Alternatively, if you prefer to not use `trust_remote_code=True` you can download [instruct_pipeline.py](https://huggingface.co/databricks/dolly-v2-3b/blob/main/instruct_pipeline.py), store it alongside your notebook, and construct the pipeline yourself from the loaded model and tokenizer: ```python import torch from instruct_pipeline import InstructionTextGenerationPipeline from transformers import AutoModelForCausalLM, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", padding_side="left") model = AutoModelForCausalLM.from_pretrained("danielpark/ko-llama-2-jindo-7b-instruct", device_map="auto", torch_dtype=torch.bfloat16) generate_text = InstructionTextGenerationPipeline(model=model, tokenizer=tokenizer) ``` ### LangChain Usage To use the pipeline with LangChain, you must set `return_full_text=True`, as LangChain expects the full text to be returned and the default for the pipeline is to only return the new text. ```python import torch from transformers import pipeline generate_text = pipeline(model="danielpark/ko-llama-2-jindo-7b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto", return_full_text=True) ``` You can create a prompt that either has only an instruction or has an instruction with context: ```python from langchain import PromptTemplate, LLMChain from langchain.llms import HuggingFacePipeline # template for an instrution with no input prompt = PromptTemplate( input_variables=["instruction"], template="{instruction}") # template for an instruction with input prompt_with_context = PromptTemplate( input_variables=["instruction", "context"], template="{instruction}\n\nInput:\n{context}") hf_pipeline = HuggingFacePipeline(pipeline=generate_text) llm_chain = LLMChain(llm=hf_pipeline, prompt=prompt) llm_context_chain = LLMChain(llm=hf_pipeline, prompt=prompt_with_context) ``` Example predicting using a simple instruction: ```python print(llm_chain.predict(instruction="Explain to me the difference between nuclear fission and fusion.").lstrip()) ``` Example predicting using an instruction with context: ```python context = """George Washington (February 22, 1732[b] - December 14, 1799) was an American military officer, statesman, and Founding Father who served as the first president of the United States from 1789 to 1797.""" print(llm_context_chain.predict(instruction="When was George Washington president?", context=context).lstrip()) ``` ### Scripts - Prepare evaluation environments: ``` # clone the repository git clone https://github.com/EleutherAI/lm-evaluation-harness.git # check out the specific commit git checkout b281b0921b636bc36ad05c0b0b0763bd6dd43463 # change to the repository directory cd lm-evaluation-harness ``` ## Ethical Issues The Jindo model has not been filtered for harmful, biased, or explicit content. As a result, outputs that do not adhere to ethical norms may be generated during use. Please exercise caution when using the model in research or practical applications. ### Ethical Considerations - There were no ethical issues involved, as we did not include the benchmark test set or the training set in the model's training process. As always, we encourage responsible and ethical use of this model. Please note that while Guanaco strives to provide accurate and helpful responses, it is still crucial to cross-verify the information from reliable sources for knowledge-based queries. ## Contact Me To contact me, you can [mail to me. [email protected]](mailto:[email protected]). ## Model Architecture ```python LlamaForCausalLM( (model): LlamaModel( (embed_tokens): Embedding(32000, 4096, padding_idx=0) (layers): ModuleList( (0-31): 32 x LlamaDecoderLayer( (self_attn): LlamaAttention( (q_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (k_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (v_proj): Linear4bit( in_features=4096, out_features=4096, bias=False (lora_dropout): ModuleDict( (default): Dropout(p=0.1, inplace=False) ) (lora_A): ModuleDict( (default): Linear(in_features=4096, out_features=64, bias=False) ) (lora_B): ModuleDict( (default): Linear(in_features=64, out_features=4096, bias=False) ) (lora_embedding_A): ParameterDict() (lora_embedding_B): ParameterDict() ) (o_proj): Linear4bit(in_features=4096, out_features=4096, bias=False) (rotary_emb): LlamaRotaryEmbedding() ) (mlp): LlamaMLP( (gate_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (up_proj): Linear4bit(in_features=4096, out_features=11008, bias=False) (down_proj): Linear4bit(in_features=11008, out_features=4096, bias=False) (act_fn): SiLUActivation() ) (input_layernorm): LlamaRMSNorm() (post_attention_layernorm): LlamaRMSNorm() ) ) (norm): LlamaRMSNorm() ) (lm_head): Linear(in_features=4096, out_features=32000, bias=False) ) ``` ## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0 ### License: The licenses of the pretrained models, llama1, and llama2, along with the datasets used, are applicable. For other datasets related to this work, guidance will be provided in the official release. The responsibility for verifying all licenses lies with the user, and the developer assumes no liability, explicit or implied, including legal responsibilities. ### Remark: - The "instruct" in the model name can be omitted, but it is used to differentiate between the backbones of llama2 for chat and general purposes. Additionally, this model is created for a specific purpose, so we plan to fine-tune it with a dataset focused on instructions. ### Naive Cost Estimation Assuming linearity without considering various variables. - 1000 simple prompts - 20 minutes of processing time - Approximately $2 in cost (based on Google Colab's 100 computing units and a single A100 GPU estimated at $10). - Required CPU RAM: 5GB (depending on the training data and dummy size) - Required VRAM: 12-13GB Time ($t$) and cost ($c$) are calculated based on the given information as follows: \[ t(n) = \frac{{20 \text{ minutes}}}{{1000 \text{ prompts}}} \times n \] \[ c(n) = \frac{{2 \text{ dollars}}}{{1000 \text{ prompts}}} \times n \] ```python def calculate_time_cost(num_prompts): total_prompts = 1000 total_time_minutes = 20 total_cost_dollars = 2 time_required = (total_time_minutes / total_prompts) * num_prompts cost_required = (total_cost_dollars / total_prompts) * num_prompts return time_required, cost_required # Example num_prompts = n time, cost = calculate_time_cost(num_prompts) print(f"Time for {num_prompts}: {time} minutes") print(f"Cost for {num_prompts}: {cost} dollar") ``` ### Chinchilla scaling laws The Chinchilla scaling laws focus on optimally scaling training compute but often we also care about inference cost. This tool follows [Harm de Vries’ blog post](https://www.harmdevries.com/post/model-size-vs-compute-overhead/) and visualizes the tradeoff between training comput and inference cost (i.e. model size).

{"datasets": ["korean-jindo-dataset.json"], "language": ["en", "ko"], "library_name": "peft", "pipeline_tag": "text-generation", "tags": ["dsdanielpark", "llama2", "instruct", "instruction", "jindo", "korean", "translation", "13b"]}

task

[ "TRANSLATION" ]

44,599

cahya/bert2bert-indonesian-summarization

cahya

summarization

[ "transformers", "pytorch", "encoder-decoder", "text2text-generation", "pipeline:summarization", "summarization", "bert2bert", "id", "dataset:id_liputan6", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:05Z

2021-01-29T11:39:42+00:00

480

4

--- datasets: - id_liputan6 language: id license: apache-2.0 tags: - pipeline:summarization - summarization - bert2bert --- # Indonesian BERT2BERT Summarization Model Finetuned BERT-base summarization model for Indonesian. ## Finetuning Corpus `bert2bert-indonesian-summarization` model is based on `cahya/bert-base-indonesian-1.5G` by [cahya](https://huggingface.co/cahya), finetuned using [id_liputan6](https://huggingface.co/datasets/id_liputan6) dataset. ## Load Finetuned Model ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") ``` ## Code Sample ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") # ARTICLE_TO_SUMMARIZE = "" # generate summary input_ids = tokenizer.encode(ARTICLE_TO_SUMMARIZE, return_tensors='pt') summary_ids = model.generate(input_ids, min_length=20, max_length=80, num_beams=10, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True, no_repeat_ngram_size=2, use_cache=True, do_sample = True, temperature = 0.8, top_k = 50, top_p = 0.95) summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True) print(summary_text) ``` Output: ``` ```

null

Non_BioNLP

# Indonesian BERT2BERT Summarization Model Finetuned BERT-base summarization model for Indonesian. ## Finetuning Corpus `bert2bert-indonesian-summarization` model is based on `cahya/bert-base-indonesian-1.5G` by [cahya](https://huggingface.co/cahya), finetuned using [id_liputan6](https://huggingface.co/datasets/id_liputan6) dataset. ## Load Finetuned Model ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") ``` ## Code Sample ```python from transformers import BertTokenizer, EncoderDecoderModel tokenizer = BertTokenizer.from_pretrained("cahya/bert2bert-indonesian-summarization") tokenizer.bos_token = tokenizer.cls_token tokenizer.eos_token = tokenizer.sep_token model = EncoderDecoderModel.from_pretrained("cahya/bert2bert-indonesian-summarization") # ARTICLE_TO_SUMMARIZE = "" # generate summary input_ids = tokenizer.encode(ARTICLE_TO_SUMMARIZE, return_tensors='pt') summary_ids = model.generate(input_ids, min_length=20, max_length=80, num_beams=10, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True, no_repeat_ngram_size=2, use_cache=True, do_sample = True, temperature = 0.8, top_k = 50, top_p = 0.95) summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True) print(summary_text) ``` Output: ``` ```

{"datasets": ["id_liputan6"], "language": "id", "license": "apache-2.0", "tags": ["pipeline:summarization", "summarization", "bert2bert"]}

task

[ "SUMMARIZATION" ]

44,600

m3hrdadfi/xlmr-large-qa-sv

m3hrdadfi

question-answering

[ "transformers", "pytorch", "tf", "xlm-roberta", "question-answering", "roberta", "squad", "sv", "multilingual", "model-index", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:05Z

2021-10-12T13:50:27+00:00

111

2

--- language: - sv - multilingual metrics: - squad_v2 tags: - question-answering - xlm-roberta - roberta - squad widget: - text: Vilket datum är den svenska nationaldagen? context: Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. - text: Vad innebär helgdag i Sverige? context: Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. - text: Vilket år tillkom Sveriges nationaldag? context: Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. model-index: - name: XLM-RoBERTa large for QA (SwedishQA - 🇸🇪) results: - task: type: question-answering name: Question Answering dataset: name: SwedishQA type: swedish_qa args: sv metrics: - type: squad_v2 value: 87.97 name: Eval F1 args: max_order - type: squad_v2 value: 78.79 name: Eval Exact args: max_order --- # XLM-RoBERTa large for QA (SwedishQA - 🇸🇪) This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the [SwedishQA](https://github.com/Vottivott/building-a-swedish-qa-model) dataset. ## Hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 2.0 - mixed_precision_training: Native AMP ## Performance Evaluation results on the eval set with the official [eval script](https://worksheets.codalab.org/rest/bundles/0x6b567e1cf2e041ec80d7098f031c5c9e/contents/blob/). ### Evalset ```text "exact": 78.79554655870446, "f1": 87.97339064752278, "total": 5928 ``` ## Usage ```python from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline model_name_or_path = "m3hrdadfi/xlmr-large-qa-sv" nlp = pipeline('question-answering', model=model_name_or_path, tokenizer=model_name_or_path) context = """ Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. """ questions = [ "Vilket datum är den svenska nationaldagen?", "Vad innebär helgdag i Sverige?", "Vilket år tillkom Sveriges nationaldag?" ] kwargs = {} for question in questions: r = nlp(question=question, context=context, **kwargs) answer = " ".join([token.strip() for token in r["answer"].strip().split() if token.strip()]) print(f"{question} {answer}") ``` **Output** ```text Vilket datum är den svenska nationaldagen? 6 juni Vad innebär helgdag i Sverige? svenska flaggans dag Vilket år tillkom Sveriges nationaldag? 1983 ``` ## Authors - [Mehrdad Farahani](https://github.com/m3hrdadfi) ### Framework versions - Transformers 4.12.0.dev0 - Pytorch 1.9.1+cu111 - Datasets 1.12.1 - Tokenizers 0.10.3

null

Non_BioNLP

# XLM-RoBERTa large for QA (SwedishQA - 🇸🇪) This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the [SwedishQA](https://github.com/Vottivott/building-a-swedish-qa-model) dataset. ## Hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 2.0 - mixed_precision_training: Native AMP ## Performance Evaluation results on the eval set with the official [eval script](https://worksheets.codalab.org/rest/bundles/0x6b567e1cf2e041ec80d7098f031c5c9e/contents/blob/). ### Evalset ```text "exact": 78.79554655870446, "f1": 87.97339064752278, "total": 5928 ``` ## Usage ```python from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline model_name_or_path = "m3hrdadfi/xlmr-large-qa-sv" nlp = pipeline('question-answering', model=model_name_or_path, tokenizer=model_name_or_path) context = """ Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som "svenska flaggans dag" och det var först 1983 som dagen även fick status som nationaldag. """ questions = [ "Vilket datum är den svenska nationaldagen?", "Vad innebär helgdag i Sverige?", "Vilket år tillkom Sveriges nationaldag?" ] kwargs = {} for question in questions: r = nlp(question=question, context=context, **kwargs) answer = " ".join([token.strip() for token in r["answer"].strip().split() if token.strip()]) print(f"{question} {answer}") ``` **Output** ```text Vilket datum är den svenska nationaldagen? 6 juni Vad innebär helgdag i Sverige? svenska flaggans dag Vilket år tillkom Sveriges nationaldag? 1983 ``` ## Authors - [Mehrdad Farahani](https://github.com/m3hrdadfi) ### Framework versions - Transformers 4.12.0.dev0 - Pytorch 1.9.1+cu111 - Datasets 1.12.1 - Tokenizers 0.10.3

{"language": ["sv", "multilingual"], "metrics": ["squad_v2"], "tags": ["question-answering", "xlm-roberta", "roberta", "squad"], "widget": [{"text": "Vilket datum är den svenska nationaldagen?", "context": "Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som \"svenska flaggans dag\" och det var först 1983 som dagen även fick status som nationaldag."}, {"text": "Vad innebär helgdag i Sverige?", "context": "Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som \"svenska flaggans dag\" och det var först 1983 som dagen även fick status som nationaldag."}, {"text": "Vilket år tillkom Sveriges nationaldag?", "context": "Sveriges nationaldag och svenska flaggans dag firas den 6 juni varje år och är en helgdag i Sverige. Tidigare firades 6 juni enbart som \"svenska flaggans dag\" och det var först 1983 som dagen även fick status som nationaldag."}], "model-index": [{"name": "XLM-RoBERTa large for QA (SwedishQA - 🇸🇪)", "results": [{"task": {"type": "question-answering", "name": "Question Answering"}, "dataset": {"name": "SwedishQA", "type": "swedish_qa", "args": "sv"}, "metrics": [{"type": "squad_v2", "value": 87.97, "name": "Eval F1", "args": "max_order"}, {"type": "squad_v2", "value": 78.79, "name": "Eval Exact", "args": "max_order"}]}]}]}

task

[ "QUESTION_ANSWERING" ]

44,601

Helsinki-NLP/opus-mt-swc-sv

Helsinki-NLP

translation

[ "transformers", "pytorch", "tf", "marian", "text2text-generation", "translation", "swc", "sv", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:04Z

2023-08-16T12:06:38+00:00

372

0

--- license: apache-2.0 tags: - translation --- ### opus-mt-swc-sv * source languages: swc * target languages: sv * OPUS readme: [swc-sv](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/swc-sv/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.zip) * test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.test.txt) * test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.eval.txt) ## Benchmarks | testset | BLEU | chr-F | |-----------------------|-------|-------| | JW300.swc.sv | 30.7 | 0.495 |

null

Non_BioNLP

### opus-mt-swc-sv * source languages: swc * target languages: sv * OPUS readme: [swc-sv](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/swc-sv/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.zip) * test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.test.txt) * test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/swc-sv/opus-2020-01-16.eval.txt) ## Benchmarks | testset | BLEU | chr-F | |-----------------------|-------|-------| | JW300.swc.sv | 30.7 | 0.495 |

{"license": "apache-2.0", "tags": ["translation"]}

task

[ "TRANSLATION" ]

44,602

mlninja-dev/distilbert-base-uncased-finetuned-emotion

mlninja-dev

text-classification

[ "transformers", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "dataset:emotion", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-11-10T07:43:27Z

2023-11-10T08:13:53+00:00

8

0

--- base_model: distilbert-base-uncased datasets: - emotion license: apache-2.0 metrics: - accuracy - f1 - precision - recall tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion config: split split: validation args: split metrics: - type: accuracy value: 0.9265 name: Accuracy - type: f1 value: 0.9265091764064305 name: F1 - type: precision value: 0.9265685739545657 name: Precision - type: recall value: 0.9265 name: Recall ---  # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2114 - Accuracy: 0.9265 - F1: 0.9265 - Precision: 0.9266 - Recall: 0.9265 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:| | 0.8163 | 1.0 | 250 | 0.3191 | 0.9055 | 0.9045 | 0.9083 | 0.9055 | | 0.2509 | 2.0 | 500 | 0.2114 | 0.9265 | 0.9265 | 0.9266 | 0.9265 | ### Framework versions - Transformers 4.35.0 - Pytorch 2.1.0 - Datasets 2.14.6 - Tokenizers 0.14.1

null

Non_BioNLP

# distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2114 - Accuracy: 0.9265 - F1: 0.9265 - Precision: 0.9266 - Recall: 0.9265 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:| | 0.8163 | 1.0 | 250 | 0.3191 | 0.9055 | 0.9045 | 0.9083 | 0.9055 | | 0.2509 | 2.0 | 500 | 0.2114 | 0.9265 | 0.9265 | 0.9266 | 0.9265 | ### Framework versions - Transformers 4.35.0 - Pytorch 2.1.0 - Datasets 2.14.6 - Tokenizers 0.14.1

{"base_model": "distilbert-base-uncased", "datasets": ["emotion"], "license": "apache-2.0", "metrics": ["accuracy", "f1", "precision", "recall"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert-base-uncased-finetuned-emotion", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "config": "split", "split": "validation", "args": "split"}, "metrics": [{"type": "accuracy", "value": 0.9265, "name": "Accuracy"}, {"type": "f1", "value": 0.9265091764064305, "name": "F1"}, {"type": "precision", "value": 0.9265685739545657, "name": "Precision"}, {"type": "recall", "value": 0.9265, "name": "Recall"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,603

prithivMLmods/Llama-Chat-Summary-3.2-3B-GGUF

prithivMLmods

text-generation

[ "transformers", "gguf", "llama", "safetensors", "ollama", "llama-cpp", "text-generation-inference", "chat-summary", "text-generation", "en", "dataset:prithivMLmods/Context-Based-Chat-Summary-Plus", "base_model:prithivMLmods/Llama-Chat-Summary-3.2-3B", "base_model:quantized:prithivMLmods/Llama-Chat-Summary-3.2-3B", "license:creativeml-openrail-m", "endpoints_compatible", "region:us", "conversational" ]

2024-12-17T04:34:28Z

2024-12-17T09:46:03+00:00

941

3

--- base_model: - prithivMLmods/Llama-Chat-Summary-3.2-3B datasets: - prithivMLmods/Context-Based-Chat-Summary-Plus language: - en library_name: transformers license: creativeml-openrail-m pipeline_tag: text-generation tags: - safetensors - ollama - llama-cpp - text-generation-inference - chat-summary --- ### **Llama-Chat-Summary-3.2-3B: Context-Aware Summarization Model** **Llama-Chat-Summary-3.2-3B** is a fine-tuned model designed for generating **context-aware summaries** of long conversational or text-based inputs. Built on the **meta-llama/Llama-3.2-3B-Instruct** foundation, this model is optimized to process structured and unstructured conversational data for summarization tasks. | **File Name** | **Size** | **Description** | **Upload Status** | |--------------------------------------------|------------------|--------------------------------------------------|-------------------| | `.gitattributes` | 1.81 kB | Git LFS tracking configuration. | Uploaded | | `Llama-Chat-Summary-3.2-3B.F16.gguf` | 6.43 GB | Full precision (F16) GGUF model file. | Uploaded (LFS) | | `Llama-Chat-Summary-3.2-3B.Q4_K_M.gguf` | 2.02 GB | Quantized Q4_K_M GGUF model file. | Uploaded (LFS) | | `Llama-Chat-Summary-3.2-3B.Q5_K_M.gguf` | 2.32 GB | Quantized Q5_K_M GGUF model file. | Uploaded (LFS) | | `Llama-Chat-Summary-3.2-3B.Q8_0.gguf` | 3.42 GB | Quantized Q8_0 GGUF model file. | Uploaded (LFS) | | `Modelfile` | 2.03 kB | Model configuration or build script file. | Uploaded | | `README.md` | 42 Bytes | Minimal commit message placeholder. | Uploaded | | `config.json` | 31 Bytes | Model metadata and configuration. | Uploaded | ### **Key Features** 1. **Conversation Summarization:** - Generates concise and meaningful summaries of long chats, discussions, or threads. 2. **Context Preservation:** - Maintains critical points, ensuring important details aren't omitted. 3. **Text Summarization:** - Works beyond chats; supports summarizing articles, documents, or reports. 4. **Fine-Tuned Efficiency:** - Trained with *Context-Based-Chat-Summary-Plus* dataset for accurate summarization of chat and conversational data. --- ### **Training Details** - **Base Model:** [meta-llama/Llama-3.2-3B-Instruct](#) - **Fine-Tuning Dataset:** [prithivMLmods/Context-Based-Chat-Summary-Plus](#) - Contains **98.4k** structured and unstructured conversations, summaries, and contextual inputs for robust training. --- ### **Applications** 1. **Customer Support Logs:** - Summarize chat logs or support tickets for insights and reporting. 2. **Meeting Notes:** - Generate concise summaries of meeting transcripts. 3. **Document Summarization:** - Create short summaries for lengthy reports or articles. 4. **Content Generation Pipelines:** - Automate summarization for newsletters, blogs, or email digests. 5. **Context Extraction for AI Systems:** - Preprocess chat or conversation logs for downstream AI applications. #### **Load the Model** ```python from transformers import AutoModelForCausalLM, AutoTokenizer model_name = "prithivMLmods/Llama-Chat-Summary-3.2-3B" tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained(model_name) ``` #### **Generate a Summary** ```python prompt = """ Summarize the following conversation: User1: Hey, I need help with my order. It hasn't arrived yet. User2: I'm sorry to hear that. Can you provide your order number? User1: Sure, it's 12345. User2: Let me check... It seems there was a delay. It should arrive tomorrow. User1: Okay, thank you! """ inputs = tokenizer(prompt, return_tensors="pt") outputs = model.generate(**inputs, max_length=100, temperature=0.7) summary = tokenizer.decode(outputs[0], skip_special_tokens=True) print("Summary:", summary) ``` --- ### **Expected Output** **"The user reported a delayed order (12345), and support confirmed it will arrive tomorrow."** --- ### **Deployment Notes** - **Serverless API:** This model currently lacks sufficient usage for serverless endpoints. Use **dedicated endpoints** for deployment. - **Performance Requirements:** - GPU with sufficient memory (recommended for large models). - Optimization techniques like quantization can improve efficiency for inference. --- # Run with Ollama [ Ollama Run ] ## Overview Ollama is a powerful tool that allows you to run machine learning models effortlessly. This guide will help you download, install, and run your own GGUF models in just a few minutes. ## Table of Contents - [Download and Install Ollama](#download-and-install-ollama) - [Steps to Run GGUF Models](#steps-to-run-gguf-models) - [1. Create the Model File](#1-create-the-model-file) - [2. Add the Template Command](#2-add-the-template-command) - [3. Create and Patch the Model](#3-create-and-patch-the-model) - [Running the Model](#running-the-model) - [Sample Usage](#sample-usage) ## Download and Install Ollama🦙 To get started, download Ollama from [https://ollama.com/download](https://ollama.com/download) and install it on your Windows or Mac system. ## Steps to Run GGUF Models ### 1. Create the Model File First, create a model file and name it appropriately. For example, you can name your model file `metallama`. ### 2. Add the Template Command In your model file, include a `FROM` line that specifies the base model file you want to use. For instance: ```bash FROM Llama-3.2-1B.F16.gguf ``` Ensure that the model file is in the same directory as your script. ### 3. Create and Patch the Model Open your terminal and run the following command to create and patch your model: ```bash ollama create metallama -f ./metallama ``` Once the process is successful, you will see a confirmation message. To verify that the model was created successfully, you can list all models with: ```bash ollama list ``` Make sure that `metallama` appears in the list of models. --- ## Running the Model To run your newly created model, use the following command in your terminal: ```bash ollama run metallama ``` ### Sample Usage / Test In the command prompt, you can execute: ```bash D:\>ollama run metallama ``` You can interact with the model like this: ```plaintext >>> write a mini passage about space x Space X, the private aerospace company founded by Elon Musk, is revolutionizing the field of space exploration. With its ambitious goals to make humanity a multi-planetary species and establish a sustainable human presence in the cosmos, Space X has become a leading player in the industry. The company's spacecraft, like the Falcon 9, have demonstrated remarkable capabilities, allowing for the transport of crews and cargo into space with unprecedented efficiency. As technology continues to advance, the possibility of establishing permanent colonies on Mars becomes increasingly feasible, thanks in part to the success of reusable rockets that can launch multiple times without sustaining significant damage. The journey towards becoming a multi-planetary species is underway, and Space X plays a pivotal role in pushing the boundaries of human exploration and settlement. ``` --- ## Conclusion With these simple steps, you can easily download, install, and run your own models using Ollama. Whether you're exploring the capabilities of Llama or building your own custom models, Ollama makes it accessible and efficient. - This README provides clear instructions and structured information to help users navigate the process of using Ollama effectively. Adjust any sections as needed based on your specific requirements or additional details you may want to include. ---

null

Non_BioNLP

### **Llama-Chat-Summary-3.2-3B: Context-Aware Summarization Model** **Llama-Chat-Summary-3.2-3B** is a fine-tuned model designed for generating **context-aware summaries** of long conversational or text-based inputs. Built on the **meta-llama/Llama-3.2-3B-Instruct** foundation, this model is optimized to process structured and unstructured conversational data for summarization tasks. | **File Name** | **Size** | **Description** | **Upload Status** | |--------------------------------------------|------------------|--------------------------------------------------|-------------------| | `.gitattributes` | 1.81 kB | Git LFS tracking configuration. | Uploaded | | `Llama-Chat-Summary-3.2-3B.F16.gguf` | 6.43 GB | Full precision (F16) GGUF model file. | Uploaded (LFS) | | `Llama-Chat-Summary-3.2-3B.Q4_K_M.gguf` | 2.02 GB | Quantized Q4_K_M GGUF model file. | Uploaded (LFS) | | `Llama-Chat-Summary-3.2-3B.Q5_K_M.gguf` | 2.32 GB | Quantized Q5_K_M GGUF model file. | Uploaded (LFS) | | `Llama-Chat-Summary-3.2-3B.Q8_0.gguf` | 3.42 GB | Quantized Q8_0 GGUF model file. | Uploaded (LFS) | | `Modelfile` | 2.03 kB | Model configuration or build script file. | Uploaded | | `README.md` | 42 Bytes | Minimal commit message placeholder. | Uploaded | | `config.json` | 31 Bytes | Model metadata and configuration. | Uploaded | ### **Key Features** 1. **Conversation Summarization:** - Generates concise and meaningful summaries of long chats, discussions, or threads. 2. **Context Preservation:** - Maintains critical points, ensuring important details aren't omitted. 3. **Text Summarization:** - Works beyond chats; supports summarizing articles, documents, or reports. 4. **Fine-Tuned Efficiency:** - Trained with *Context-Based-Chat-Summary-Plus* dataset for accurate summarization of chat and conversational data. --- ### **Training Details** - **Base Model:** [meta-llama/Llama-3.2-3B-Instruct](#) - **Fine-Tuning Dataset:** [prithivMLmods/Context-Based-Chat-Summary-Plus](#) - Contains **98.4k** structured and unstructured conversations, summaries, and contextual inputs for robust training. --- ### **Applications** 1. **Customer Support Logs:** - Summarize chat logs or support tickets for insights and reporting. 2. **Meeting Notes:** - Generate concise summaries of meeting transcripts. 3. **Document Summarization:** - Create short summaries for lengthy reports or articles. 4. **Content Generation Pipelines:** - Automate summarization for newsletters, blogs, or email digests. 5. **Context Extraction for AI Systems:** - Preprocess chat or conversation logs for downstream AI applications. #### **Load the Model** ```python from transformers import AutoModelForCausalLM, AutoTokenizer model_name = "prithivMLmods/Llama-Chat-Summary-3.2-3B" tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained(model_name) ``` #### **Generate a Summary** ```python prompt = """ Summarize the following conversation: User1: Hey, I need help with my order. It hasn't arrived yet. User2: I'm sorry to hear that. Can you provide your order number? User1: Sure, it's 12345. User2: Let me check... It seems there was a delay. It should arrive tomorrow. User1: Okay, thank you! """ inputs = tokenizer(prompt, return_tensors="pt") outputs = model.generate(**inputs, max_length=100, temperature=0.7) summary = tokenizer.decode(outputs[0], skip_special_tokens=True) print("Summary:", summary) ``` --- ### **Expected Output** **"The user reported a delayed order (12345), and support confirmed it will arrive tomorrow."** --- ### **Deployment Notes** - **Serverless API:** This model currently lacks sufficient usage for serverless endpoints. Use **dedicated endpoints** for deployment. - **Performance Requirements:** - GPU with sufficient memory (recommended for large models). - Optimization techniques like quantization can improve efficiency for inference. --- # Run with Ollama [ Ollama Run ] ## Overview Ollama is a powerful tool that allows you to run machine learning models effortlessly. This guide will help you download, install, and run your own GGUF models in just a few minutes. ## Table of Contents - [Download and Install Ollama](#download-and-install-ollama) - [Steps to Run GGUF Models](#steps-to-run-gguf-models) - [1. Create the Model File](#1-create-the-model-file) - [2. Add the Template Command](#2-add-the-template-command) - [3. Create and Patch the Model](#3-create-and-patch-the-model) - [Running the Model](#running-the-model) - [Sample Usage](#sample-usage) ## Download and Install Ollama🦙 To get started, download Ollama from [https://ollama.com/download](https://ollama.com/download) and install it on your Windows or Mac system. ## Steps to Run GGUF Models ### 1. Create the Model File First, create a model file and name it appropriately. For example, you can name your model file `metallama`. ### 2. Add the Template Command In your model file, include a `FROM` line that specifies the base model file you want to use. For instance: ```bash FROM Llama-3.2-1B.F16.gguf ``` Ensure that the model file is in the same directory as your script. ### 3. Create and Patch the Model Open your terminal and run the following command to create and patch your model: ```bash ollama create metallama -f ./metallama ``` Once the process is successful, you will see a confirmation message. To verify that the model was created successfully, you can list all models with: ```bash ollama list ``` Make sure that `metallama` appears in the list of models. --- ## Running the Model To run your newly created model, use the following command in your terminal: ```bash ollama run metallama ``` ### Sample Usage / Test In the command prompt, you can execute: ```bash D:\>ollama run metallama ``` You can interact with the model like this: ```plaintext >>> write a mini passage about space x Space X, the private aerospace company founded by Elon Musk, is revolutionizing the field of space exploration. With its ambitious goals to make humanity a multi-planetary species and establish a sustainable human presence in the cosmos, Space X has become a leading player in the industry. The company's spacecraft, like the Falcon 9, have demonstrated remarkable capabilities, allowing for the transport of crews and cargo into space with unprecedented efficiency. As technology continues to advance, the possibility of establishing permanent colonies on Mars becomes increasingly feasible, thanks in part to the success of reusable rockets that can launch multiple times without sustaining significant damage. The journey towards becoming a multi-planetary species is underway, and Space X plays a pivotal role in pushing the boundaries of human exploration and settlement. ``` --- ## Conclusion With these simple steps, you can easily download, install, and run your own models using Ollama. Whether you're exploring the capabilities of Llama or building your own custom models, Ollama makes it accessible and efficient. - This README provides clear instructions and structured information to help users navigate the process of using Ollama effectively. Adjust any sections as needed based on your specific requirements or additional details you may want to include. ---

{"base_model": ["prithivMLmods/Llama-Chat-Summary-3.2-3B"], "datasets": ["prithivMLmods/Context-Based-Chat-Summary-Plus"], "language": ["en"], "library_name": "transformers", "license": "creativeml-openrail-m", "pipeline_tag": "text-generation", "tags": ["safetensors", "ollama", "llama-cpp", "text-generation-inference", "chat-summary"]}

task

[ "SUMMARIZATION" ]

44,604

cnmoro/ptt5-base-ptbr-summarization

cnmoro

summarization

[ "transformers", "pytorch", "t5", "text2text-generation", "summarization", "pt", "license:mit", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2023-11-09T14:33:11Z

2023-11-10T12:54:28+00:00

651

2

--- language: - pt license: mit tags: - summarization ---

null

Non_BioNLP

{"language": ["pt"], "license": "mit", "tags": ["summarization"]}

task

[ "SUMMARIZATION" ]

44,605

deinon-daemon/axolotl-13b-chat-qlora-dev

deinon-daemon

text-generation

[ "transformers", "pytorch", "llama", "text-generation", "en", "license:llama2", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2023-07-29T11:48:10Z

2023-07-29T14:38:16+00:00

10

0

--- language: en license: llama2 task_categories: - conversational - question-answering - zero-shot-answering - summarization - text-generation - text2text-generation pretty_name: Axolotl-13B-chat --- <img src="https://huggingface.co/deinon-daemon/axolotl-13b-chat-qlora-dev/resolve/main/axolotl_img/a14a0db4-9c8c-4283-8d24-21eb8f6210b1.png" alt="¡ Say hi to Axolotl !" style="float: left; margin-right: 10px;" /> Say hello to axolotl: a small-is-powerful instruct-tuned chat model! This is my second build ever in the fine tuning world. It was hacked in about 48hrs, and was executed entirely on one colab kernel for ~8-9hrs last night (07/29/23) ... enjoy! Test run of Llama-2-13b-chat-hf fine tuned using recently popularized quantized PEFT approach: used Bitsandbytes, --bf16, QLORA, Flash Attn w/ einops and ninja Ampere optimizations, 1 Nvidia A100 GPU for ~9hrs. Fine tuned for 3 epochs on a 40k slice of the Open-Orca dataset, which I postprocessed, added some self-collected contextual qa chat data to, and templated to yield a standard chat instruct prompt format for all examples. Benchmarks at least as good (if not slightly better) than other fine tuned llama/alpaca/guanaco/vicuna models of this scale. The real evaulation/benchmarking is still to come, however, specifically against stabilityai/StableBeluga13B, which seems to be the most popular example of Llama-2 + Open-Orca to date. This is simply a proof of concept (hence the dev tag) -- come back later once we've realeased a model for production.

null

Non_BioNLP

<img src="https://huggingface.co/deinon-daemon/axolotl-13b-chat-qlora-dev/resolve/main/axolotl_img/a14a0db4-9c8c-4283-8d24-21eb8f6210b1.png" alt="¡ Say hi to Axolotl !" style="float: left; margin-right: 10px;" /> Say hello to axolotl: a small-is-powerful instruct-tuned chat model! This is my second build ever in the fine tuning world. It was hacked in about 48hrs, and was executed entirely on one colab kernel for ~8-9hrs last night (07/29/23) ... enjoy! Test run of Llama-2-13b-chat-hf fine tuned using recently popularized quantized PEFT approach: used Bitsandbytes, --bf16, QLORA, Flash Attn w/ einops and ninja Ampere optimizations, 1 Nvidia A100 GPU for ~9hrs. Fine tuned for 3 epochs on a 40k slice of the Open-Orca dataset, which I postprocessed, added some self-collected contextual qa chat data to, and templated to yield a standard chat instruct prompt format for all examples. Benchmarks at least as good (if not slightly better) than other fine tuned llama/alpaca/guanaco/vicuna models of this scale. The real evaulation/benchmarking is still to come, however, specifically against stabilityai/StableBeluga13B, which seems to be the most popular example of Llama-2 + Open-Orca to date. This is simply a proof of concept (hence the dev tag) -- come back later once we've realeased a model for production.

{"language": "en", "license": "llama2", "task_categories": ["conversational", "question-answering", "zero-shot-answering", "summarization", "text-generation", "text2text-generation"], "pretty_name": "Axolotl-13B-chat"}

task

[ "SUMMARIZATION" ]

44,606

lordtt13/blenderbot_small-news

lordtt13

text2text-generation

[ "transformers", "pytorch", "tf", "safetensors", "blenderbot-small", "text2text-generation", "en", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:05Z

2023-07-09T15:28:39+00:00

21

0

--- language: en --- ## BlenderBotSmall-News: Small version of a state-of-the-art open source chatbot, trained on custom summaries ### Details of BlenderBotSmall The **BlenderBotSmall** model was presented in [A state-of-the-art open source chatbot](https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot/) by *Facebook AI* and here are it's details: - Facebook AI has built and open-sourced BlenderBot, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators. - The culmination of years of research in conversational AI, this is the first chatbot to blend a diverse set of conversational skills — including empathy, knowledge, and personality — together in one system. - We achieved this milestone through a new chatbot recipe that includes improved decoding techniques, novel blending of skills, and a model with 9.4 billion parameters, which is 3.6x more than the largest existing system. ### Details of the downstream task (Summarization) - Dataset 📚 A custom dataset was used, which was hand prepared by [SmokeTrees Digital](https://github.com/smoke-trees) AI engineers. This data contains long texts and summaries. ### Model training The training script is present [here](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb). ### Pipelining the Model ```python model = transformers.BlenderbotSmallForConditionalGeneration.from_pretrained('lordtt13/blenderbot_small-news') tokenizer = transformers.BlenderbotSmallTokenizer.from_pretrained("lordtt13/blenderbot_small-news") nlp_fill = transformers.pipeline('summarization', model = model, tokenizer = tokenizer) nlp_fill('The CBI on Saturday booked four former officials of Syndicate Bank and six others for cheating, forgery, criminal conspiracy and causing ₹209 crore loss to the state-run bank. The accused had availed home loans and credit from Syndicate Bank on the basis of forged and fabricated documents. These funds were fraudulently transferred to the companies owned by the accused persons.', min_length=5, max_length=40) # Output: # [{'summary_text': 'marize: the cbi booked four former officials of syndicate bank and six others for cheating , forgery , criminal conspiracy and causing 209 crore loss to the staterun bank'}] ``` > Created by [Tanmay Thakur](https://github.com/lordtt13) | [LinkedIn](https://www.linkedin.com/in/tanmay-thakur-6bb5a9154/)

null

Non_BioNLP

## BlenderBotSmall-News: Small version of a state-of-the-art open source chatbot, trained on custom summaries ### Details of BlenderBotSmall The **BlenderBotSmall** model was presented in [A state-of-the-art open source chatbot](https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot/) by *Facebook AI* and here are it's details: - Facebook AI has built and open-sourced BlenderBot, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators. - The culmination of years of research in conversational AI, this is the first chatbot to blend a diverse set of conversational skills — including empathy, knowledge, and personality — together in one system. - We achieved this milestone through a new chatbot recipe that includes improved decoding techniques, novel blending of skills, and a model with 9.4 billion parameters, which is 3.6x more than the largest existing system. ### Details of the downstream task (Summarization) - Dataset 📚 A custom dataset was used, which was hand prepared by [SmokeTrees Digital](https://github.com/smoke-trees) AI engineers. This data contains long texts and summaries. ### Model training The training script is present [here](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb). ### Pipelining the Model ```python model = transformers.BlenderbotSmallForConditionalGeneration.from_pretrained('lordtt13/blenderbot_small-news') tokenizer = transformers.BlenderbotSmallTokenizer.from_pretrained("lordtt13/blenderbot_small-news") nlp_fill = transformers.pipeline('summarization', model = model, tokenizer = tokenizer) nlp_fill('The CBI on Saturday booked four former officials of Syndicate Bank and six others for cheating, forgery, criminal conspiracy and causing ₹209 crore loss to the state-run bank. The accused had availed home loans and credit from Syndicate Bank on the basis of forged and fabricated documents. These funds were fraudulently transferred to the companies owned by the accused persons.', min_length=5, max_length=40) # Output: # [{'summary_text': 'marize: the cbi booked four former officials of syndicate bank and six others for cheating , forgery , criminal conspiracy and causing 209 crore loss to the staterun bank'}] ``` > Created by [Tanmay Thakur](https://github.com/lordtt13) | [LinkedIn](https://www.linkedin.com/in/tanmay-thakur-6bb5a9154/)

{"language": "en"}

task

[ "SUMMARIZATION" ]

44,607

KameliaZaman/French-to-English-Translation

KameliaZaman

translation

[ "transformers", "translation", "en", "fr", "license:mit", "endpoints_compatible", "region:us" ]

2024-04-08T04:04:45Z

2024-04-13T04:55:49+00:00

10

0

--- language: - en - fr library_name: transformers license: mit metrics: - bleu pipeline_tag: translation --- <a name="readme-top"></a> <div align="center"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/logo.jpg" alt="Logo" width="100" height="100"> <h3 align="center">French to English Machine Translation</h3> French to English language translation using sequence to sequence transformer. <a href="https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation">View Demo</a> </div>  <details> <summary>Table of Contents</summary> <ol> <li> <a href="#about-the-project">About The Project</a> <ul> <li><a href="#built-with">Built With</a></li> </ul> </li> <li> <a href="#getting-started">Getting Started</a> <ul> <li><a href="#dependencies">Dependencies</a></li> <li><a href="#installation">Installation</a></li> </ul> </li> <li><a href="#usage">Usage</a></li> <li><a href="#contributing">Contributing</a></li> <li><a href="#license">License</a></li> <li><a href="#contact">Contact</a></li> </ol> </details>  ## About The Project <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> This project aims to develop a machine translation system for translating French text into English. The system utilizes state-of-the-art neural network architectures and techniques in natural language processing (NLP) to accurately translate French sentences into their corresponding English equivalents. (<a href="#readme-top">back to top</a>) ### Built With * [![Python][Python]][Python-url] * [![TensorFlow][TensorFlow]][TensorFlow-url] * [![Keras][Keras]][Keras-url] * [![NumPy][NumPy]][NumPy-url] * [![Pandas][Pandas]][Pandas-url] (<a href="#readme-top">back to top</a>)  ## Getting Started Please follow these simple steps to setup this project locally. ### Dependencies Here are the list all libraries, packages and other dependencies that need to be installed to run this project. For example, this is how you would list them: * TensorFlow 2.16.1 ```sh conda install -c conda-forge tensorflow ``` * Keras 2.15.0 ```sh conda install -c conda-forge keras ``` * Gradio 4.24.0 ```sh conda install -c conda-forge gradio ``` * NumPy 1.26.4 ```sh conda install -c conda-forge numpy ``` ### Alternative: Export Environment Alternatively, clone the project repository, install it and have all dependencies needed. ```sh conda env export > requirements.txt ``` Recreate it using: ```sh conda env create -f requirements.txt ``` ### Installation ```sh # clone project git clone https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main # go inside the project directory cd French-to-English-Translation # install the required packages pip install -r requirements.txt # run the gradio app python app.py ``` (<a href="#readme-top">back to top</a>)  ## Usage #### Dataset Dataset is from "https://www.kaggle.com/datasets/devicharith/language-translation-englishfrench" which contains 2 columns where one column has english words/sentences and the other one has french words/sentence #### Model Architecture The model architecture consists of an Encoder-Decoder Long Short-Term Memory network with an embedding layer. It was built on a Neural Machine Translation architecture where sequence-to-sequence framework with attention mechanisms was applied. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/arch.png" alt="Logo" width="500" height="500"> #### Data Preparation - The parallel corpus containing French and English sentences is preprocessed. - Text is tokenized and converted into numerical representations suitable for input to the neural network. #### Model Training - The sequence-to-sequence model is constructed, comprising an encoder and decoder. - Training data is fed into the model, and parameters are optimized using backpropagation and gradient descent algorithms. ```sh def create_model(src_vocab, tar_vocab, src_timesteps, tar_timesteps, n_units): # Create the model model = Sequential() model.add(Embedding(src_vocab_size, n_units, input_length=src_length, mask_zero=True)) model.add(LSTM(n_units)) model.add(RepeatVector(tar_timesteps)) model.add(LSTM(n_units, return_sequences=True)) model.add(TimeDistributed(Dense(tar_vocab, activation='softmax'))) return model model = create_model(src_vocab_size, tar_vocab_size, src_length, tar_length, 256) model.compile(optimizer='adam', loss='categorical_crossentropy') history = model.fit(trainX, trainY, epochs=20, batch_size=64, validation_split=0.1, verbose=1, callbacks=[ EarlyStopping( monitor='val_loss', patience=10, restore_best_weights=True ) ]) ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_loss.png" alt="Logo" width="500" height="500"> #### Model Evaluation - The trained model is evaluated on the test set to measure its accuracy. - Metrics such as BLEU score has been used to quantify the quality of translations. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_acc.png" alt="Logo" width="500" height="500"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/test_acc.png" alt="Logo" width="500" height="500"> #### Deployment - Gradio is utilized for deploying the trained model. - Users can input a French text, and the model will translate it to English. ```sh import string import re from unicodedata import normalize import numpy as np from keras.preprocessing.text import Tokenizer from keras.preprocessing.sequence import pad_sequences from keras.utils import to_categorical from keras.models import Sequential,load_model from keras.layers import LSTM,Dense,Embedding,RepeatVector,TimeDistributed from keras.callbacks import EarlyStopping from nltk.translate.bleu_score import corpus_bleu import pandas as pd from string import punctuation import matplotlib.pyplot as plt from IPython.display import Markdown, display import gradio as gr import tensorflow as tf from tensorflow.keras.models import load_model total_sentences = 10000 dataset = pd.read_csv("./eng_-french.csv", nrows = total_sentences) def clean(string): # Clean the string string = string.replace("\u202f"," ") # Replace no-break space with space string = string.lower() # Delete the punctuation and the numbers for p in punctuation + "«»" + "0123456789": string = string.replace(p," ") string = re.sub('\s+',' ', string) string = string.strip() return string dataset = dataset.sample(frac=1, random_state=0) dataset["English words/sentences"] = dataset["English words/sentences"].apply(lambda x: clean(x)) dataset["French words/sentences"] = dataset["French words/sentences"].apply(lambda x: clean(x)) dataset = dataset.values dataset = dataset[:total_sentences] source_str, target_str = "French", "English" idx_src, idx_tar = 1, 0 def create_tokenizer(lines): # fit a tokenizer tokenizer = Tokenizer() tokenizer.fit_on_texts(lines) return tokenizer def max_len(lines): # max sentence length return max(len(line.split()) for line in lines) def encode_sequences(tokenizer, length, lines): # encode and pad sequences X = tokenizer.texts_to_sequences(lines) # integer encode sequences X = pad_sequences(X, maxlen=length, padding='post') # pad sequences with 0 values return X def word_for_id(integer, tokenizer): # map an integer to a word for word, index in tokenizer.word_index.items(): if index == integer: return word return None def predict_seq(model, tokenizer, source): # generate target from a source sequence prediction = model.predict(source, verbose=0)[0] integers = [np.argmax(vector) for vector in prediction] target = list() for i in integers: word = word_for_id(i, tokenizer) if word is None: break target.append(word) return ' '.join(target) src_tokenizer = create_tokenizer(dataset[:, idx_src]) src_vocab_size = len(src_tokenizer.word_index) + 1 src_length = max_len(dataset[:, idx_src]) tar_tokenizer = create_tokenizer(dataset[:, idx_tar]) model = load_model('./french_to_english_translator.h5') def translate_french_english(french_sentence): # Clean the input sentence french_sentence = clean(french_sentence) # Tokenize and pad the input sentence input_sequence = encode_sequences(src_tokenizer, src_length, [french_sentence]) # Generate the translation english_translation = predict_seq(model, tar_tokenizer, input_sequence) return english_translation gr.Interface( fn=translate_french_english, inputs="text", outputs="text", title="French to English Translator", description="Translate French sentences to English." ).launch() ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> (<a href="#readme-top">back to top</a>)  ## Contributing Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are **greatly appreciated**. If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again! 1. Fork the Project 2. Create your Feature Branch (`git checkout -b feature/AmazingFeature`) 3. Commit your Changes (`git commit -m 'Add some AmazingFeature'`) 4. Push to the Branch (`git push origin feature/AmazingFeature`) 5. Open a Pull Request (<a href="#readme-top">back to top</a>)  ## License Distributed under the MIT License. See [MIT License](LICENSE) for more information. (<a href="#readme-top">back to top</a>)  ## Contact Kamelia Zaman Moon - [email protected] Project Link: [https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation](https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main) (<a href="#readme-top">back to top</a>) [Python]: https://img.shields.io/badge/python-3670A0?style=for-the-badge&logo=python&logoColor=ffdd54 [Python-url]: https://www.python.org/ [TensorFlow]: https://img.shields.io/badge/TensorFlow-%23FF6F00.svg?style=for-the-badge&logo=TensorFlow&logoColor=white [TensorFlow-url]: https://tensorflow.org/ [Keras]: https://img.shields.io/badge/Keras-%23D00000.svg?style=for-the-badge&logo=Keras&logoColor=white [Keras-url]: https://keras.io/ [NumPy]: https://img.shields.io/badge/numpy-%23013243.svg?style=for-the-badge&logo=numpy&logoColor=white [NumPy-url]: https://numpy.org/ [Pandas]: https://img.shields.io/badge/pandas-%23150458.svg?style=for-the-badge&logo=pandas&logoColor=white [Pandas-url]: https://pandas.pydata.org/

null

Non_BioNLP

<a name="readme-top"></a> <div align="center"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/logo.jpg" alt="Logo" width="100" height="100"> <h3 align="center">French to English Machine Translation</h3> French to English language translation using sequence to sequence transformer. <a href="https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation">View Demo</a> </div>  <details> <summary>Table of Contents</summary> <ol> <li> <a href="#about-the-project">About The Project</a> <ul> <li><a href="#built-with">Built With</a></li> </ul> </li> <li> <a href="#getting-started">Getting Started</a> <ul> <li><a href="#dependencies">Dependencies</a></li> <li><a href="#installation">Installation</a></li> </ul> </li> <li><a href="#usage">Usage</a></li> <li><a href="#contributing">Contributing</a></li> <li><a href="#license">License</a></li> <li><a href="#contact">Contact</a></li> </ol> </details>  ## About The Project <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> This project aims to develop a machine translation system for translating French text into English. The system utilizes state-of-the-art neural network architectures and techniques in natural language processing (NLP) to accurately translate French sentences into their corresponding English equivalents. (<a href="#readme-top">back to top</a>) ### Built With * [![Python][Python]][Python-url] * [![TensorFlow][TensorFlow]][TensorFlow-url] * [![Keras][Keras]][Keras-url] * [![NumPy][NumPy]][NumPy-url] * [![Pandas][Pandas]][Pandas-url] (<a href="#readme-top">back to top</a>)  ## Getting Started Please follow these simple steps to setup this project locally. ### Dependencies Here are the list all libraries, packages and other dependencies that need to be installed to run this project. For example, this is how you would list them: * TensorFlow 2.16.1 ```sh conda install -c conda-forge tensorflow ``` * Keras 2.15.0 ```sh conda install -c conda-forge keras ``` * Gradio 4.24.0 ```sh conda install -c conda-forge gradio ``` * NumPy 1.26.4 ```sh conda install -c conda-forge numpy ``` ### Alternative: Export Environment Alternatively, clone the project repository, install it and have all dependencies needed. ```sh conda env export > requirements.txt ``` Recreate it using: ```sh conda env create -f requirements.txt ``` ### Installation ```sh # clone project git clone https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main # go inside the project directory cd French-to-English-Translation # install the required packages pip install -r requirements.txt # run the gradio app python app.py ``` (<a href="#readme-top">back to top</a>)  ## Usage #### Dataset Dataset is from "https://www.kaggle.com/datasets/devicharith/language-translation-englishfrench" which contains 2 columns where one column has english words/sentences and the other one has french words/sentence #### Model Architecture The model architecture consists of an Encoder-Decoder Long Short-Term Memory network with an embedding layer. It was built on a Neural Machine Translation architecture where sequence-to-sequence framework with attention mechanisms was applied. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/arch.png" alt="Logo" width="500" height="500"> #### Data Preparation - The parallel corpus containing French and English sentences is preprocessed. - Text is tokenized and converted into numerical representations suitable for input to the neural network. #### Model Training - The sequence-to-sequence model is constructed, comprising an encoder and decoder. - Training data is fed into the model, and parameters are optimized using backpropagation and gradient descent algorithms. ```sh def create_model(src_vocab, tar_vocab, src_timesteps, tar_timesteps, n_units): # Create the model model = Sequential() model.add(Embedding(src_vocab_size, n_units, input_length=src_length, mask_zero=True)) model.add(LSTM(n_units)) model.add(RepeatVector(tar_timesteps)) model.add(LSTM(n_units, return_sequences=True)) model.add(TimeDistributed(Dense(tar_vocab, activation='softmax'))) return model model = create_model(src_vocab_size, tar_vocab_size, src_length, tar_length, 256) model.compile(optimizer='adam', loss='categorical_crossentropy') history = model.fit(trainX, trainY, epochs=20, batch_size=64, validation_split=0.1, verbose=1, callbacks=[ EarlyStopping( monitor='val_loss', patience=10, restore_best_weights=True ) ]) ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_loss.png" alt="Logo" width="500" height="500"> #### Model Evaluation - The trained model is evaluated on the test set to measure its accuracy. - Metrics such as BLEU score has been used to quantify the quality of translations. <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/train_acc.png" alt="Logo" width="500" height="500"> <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/test_acc.png" alt="Logo" width="500" height="500"> #### Deployment - Gradio is utilized for deploying the trained model. - Users can input a French text, and the model will translate it to English. ```sh import string import re from unicodedata import normalize import numpy as np from keras.preprocessing.text import Tokenizer from keras.preprocessing.sequence import pad_sequences from keras.utils import to_categorical from keras.models import Sequential,load_model from keras.layers import LSTM,Dense,Embedding,RepeatVector,TimeDistributed from keras.callbacks import EarlyStopping from nltk.translate.bleu_score import corpus_bleu import pandas as pd from string import punctuation import matplotlib.pyplot as plt from IPython.display import Markdown, display import gradio as gr import tensorflow as tf from tensorflow.keras.models import load_model total_sentences = 10000 dataset = pd.read_csv("./eng_-french.csv", nrows = total_sentences) def clean(string): # Clean the string string = string.replace("\u202f"," ") # Replace no-break space with space string = string.lower() # Delete the punctuation and the numbers for p in punctuation + "«»" + "0123456789": string = string.replace(p," ") string = re.sub('\s+',' ', string) string = string.strip() return string dataset = dataset.sample(frac=1, random_state=0) dataset["English words/sentences"] = dataset["English words/sentences"].apply(lambda x: clean(x)) dataset["French words/sentences"] = dataset["French words/sentences"].apply(lambda x: clean(x)) dataset = dataset.values dataset = dataset[:total_sentences] source_str, target_str = "French", "English" idx_src, idx_tar = 1, 0 def create_tokenizer(lines): # fit a tokenizer tokenizer = Tokenizer() tokenizer.fit_on_texts(lines) return tokenizer def max_len(lines): # max sentence length return max(len(line.split()) for line in lines) def encode_sequences(tokenizer, length, lines): # encode and pad sequences X = tokenizer.texts_to_sequences(lines) # integer encode sequences X = pad_sequences(X, maxlen=length, padding='post') # pad sequences with 0 values return X def word_for_id(integer, tokenizer): # map an integer to a word for word, index in tokenizer.word_index.items(): if index == integer: return word return None def predict_seq(model, tokenizer, source): # generate target from a source sequence prediction = model.predict(source, verbose=0)[0] integers = [np.argmax(vector) for vector in prediction] target = list() for i in integers: word = word_for_id(i, tokenizer) if word is None: break target.append(word) return ' '.join(target) src_tokenizer = create_tokenizer(dataset[:, idx_src]) src_vocab_size = len(src_tokenizer.word_index) + 1 src_length = max_len(dataset[:, idx_src]) tar_tokenizer = create_tokenizer(dataset[:, idx_tar]) model = load_model('./french_to_english_translator.h5') def translate_french_english(french_sentence): # Clean the input sentence french_sentence = clean(french_sentence) # Tokenize and pad the input sentence input_sequence = encode_sequences(src_tokenizer, src_length, [french_sentence]) # Generate the translation english_translation = predict_seq(model, tar_tokenizer, input_sequence) return english_translation gr.Interface( fn=translate_french_english, inputs="text", outputs="text", title="French to English Translator", description="Translate French sentences to English." ).launch() ``` <img src="https://huggingface.co/KameliaZaman/French-to-English-Translation/resolve/main/assets/About.png" alt="Logo" width="500" height="500"> (<a href="#readme-top">back to top</a>)  ## Contributing Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are **greatly appreciated**. If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again! 1. Fork the Project 2. Create your Feature Branch (`git checkout -b feature/AmazingFeature`) 3. Commit your Changes (`git commit -m 'Add some AmazingFeature'`) 4. Push to the Branch (`git push origin feature/AmazingFeature`) 5. Open a Pull Request (<a href="#readme-top">back to top</a>)  ## License Distributed under the MIT License. See [MIT License](LICENSE) for more information. (<a href="#readme-top">back to top</a>)  ## Contact Kamelia Zaman Moon - [email protected] Project Link: [https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation](https://huggingface.co/spaces/KameliaZaman/French-to-English-Translation/tree/main) (<a href="#readme-top">back to top</a>) [Python]: https://img.shields.io/badge/python-3670A0?style=for-the-badge&logo=python&logoColor=ffdd54 [Python-url]: https://www.python.org/ [TensorFlow]: https://img.shields.io/badge/TensorFlow-%23FF6F00.svg?style=for-the-badge&logo=TensorFlow&logoColor=white [TensorFlow-url]: https://tensorflow.org/ [Keras]: https://img.shields.io/badge/Keras-%23D00000.svg?style=for-the-badge&logo=Keras&logoColor=white [Keras-url]: https://keras.io/ [NumPy]: https://img.shields.io/badge/numpy-%23013243.svg?style=for-the-badge&logo=numpy&logoColor=white [NumPy-url]: https://numpy.org/ [Pandas]: https://img.shields.io/badge/pandas-%23150458.svg?style=for-the-badge&logo=pandas&logoColor=white [Pandas-url]: https://pandas.pydata.org/

{"language": ["en", "fr"], "library_name": "transformers", "license": "mit", "metrics": ["bleu"], "pipeline_tag": "translation"}

task

[ "TRANSLATION" ]

44,609

openai/whisper-base

openai

automatic-speech-recognition

[ "transformers", "pytorch", "tf", "jax", "safetensors", "whisper", "automatic-speech-recognition", "audio", "hf-asr-leaderboard", "en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", "no", "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su", "arxiv:2212.04356", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]

2022-09-26T06:50:46Z

2024-02-29T10:26:57+00:00

1,161,580

210

--- language: - en - zh - de - es - ru - ko - fr - ja - pt - tr - pl - ca - nl - ar - sv - it - id - hi - fi - vi - he - uk - el - ms - cs - ro - da - hu - ta - false - th - ur - hr - bg - lt - la - mi - ml - cy - sk - te - fa - lv - bn - sr - az - sl - kn - et - mk - br - eu - is - hy - ne - mn - bs - kk - sq - sw - gl - mr - pa - si - km - sn - yo - so - af - oc - ka - be - tg - sd - gu - am - yi - lo - uz - fo - ht - ps - tk - nn - mt - sa - lb - my - bo - tl - mg - as - tt - haw - ln - ha - ba - jw - su license: apache-2.0 pipeline_tag: automatic-speech-recognition tags: - audio - automatic-speech-recognition - hf-asr-leaderboard widget: - example_title: Librispeech sample 1 src: https://cdn-media.huggingface.co/speech_samples/sample1.flac - example_title: Librispeech sample 2 src: https://cdn-media.huggingface.co/speech_samples/sample2.flac model-index: - name: whisper-base results: - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (clean) type: librispeech_asr config: clean split: test args: language: en metrics: - type: wer value: 5.008769117619326 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: LibriSpeech (other) type: librispeech_asr config: other split: test args: language: en metrics: - type: wer value: 12.84936273212057 name: Test WER - task: type: automatic-speech-recognition name: Automatic Speech Recognition dataset: name: Common Voice 11.0 type: mozilla-foundation/common_voice_11_0 config: hi split: test args: language: hi metrics: - type: wer value: 131 name: Test WER --- # Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains **without** the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). **Disclaimer**: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the *same* language as the audio. For speech translation, the model predicts transcriptions to a *different* language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: | Size | Parameters | English-only | Multilingual | |----------|------------|------------------------------------------------------|-----------------------------------------------------| | tiny | 39 M | [✓](https://huggingface.co/openai/whisper-tiny.en) | [✓](https://huggingface.co/openai/whisper-tiny) | | base | 74 M | [✓](https://huggingface.co/openai/whisper-base.en) | [✓](https://huggingface.co/openai/whisper-base) | | small | 244 M | [✓](https://huggingface.co/openai/whisper-small.en) | [✓](https://huggingface.co/openai/whisper-small) | | medium | 769 M | [✓](https://huggingface.co/openai/whisper-medium.en) | [✓](https://huggingface.co/openai/whisper-medium) | | large | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large) | | large-v2 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v2) | # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<|startoftranscript|>` token 2. The second token is the language token (e.g. `<|en|>` for English) 3. The third token is the "task token". It can take one of two values: `<|transcribe|>` for speech recognition or `<|translate|>` for speech translation 4. In addition, a `<|notimestamps|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <|startoftranscript|> <|en|> <|transcribe|> <|notimestamps|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<|startoftranscript|><|en|><|transcribe|><|notimestamps|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<|startoftranscript|><|fr|><|transcribe|><|notimestamps|> Un vrai travail intéressant va enfin être mené sur ce sujet.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through *fine-tuning*. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```

null

Non_BioNLP

# Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains **without** the need for fine-tuning. Whisper was proposed in the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://arxiv.org/abs/2212.04356) by Alec Radford et al from OpenAI. The original code repository can be found [here](https://github.com/openai/whisper). **Disclaimer**: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. ## Model details Whisper is a Transformer based encoder-decoder model, also referred to as a _sequence-to-sequence_ model. It was trained on 680k hours of labelled speech data annotated using large-scale weak supervision. The models were trained on either English-only data or multilingual data. The English-only models were trained on the task of speech recognition. The multilingual models were trained on both speech recognition and speech translation. For speech recognition, the model predicts transcriptions in the *same* language as the audio. For speech translation, the model predicts transcriptions to a *different* language to the audio. Whisper checkpoints come in five configurations of varying model sizes. The smallest four are trained on either English-only or multilingual data. The largest checkpoints are multilingual only. All ten of the pre-trained checkpoints are available on the [Hugging Face Hub](https://huggingface.co/models?search=openai/whisper). The checkpoints are summarised in the following table with links to the models on the Hub: | Size | Parameters | English-only | Multilingual | |----------|------------|------------------------------------------------------|-----------------------------------------------------| | tiny | 39 M | [✓](https://huggingface.co/openai/whisper-tiny.en) | [✓](https://huggingface.co/openai/whisper-tiny) | | base | 74 M | [✓](https://huggingface.co/openai/whisper-base.en) | [✓](https://huggingface.co/openai/whisper-base) | | small | 244 M | [✓](https://huggingface.co/openai/whisper-small.en) | [✓](https://huggingface.co/openai/whisper-small) | | medium | 769 M | [✓](https://huggingface.co/openai/whisper-medium.en) | [✓](https://huggingface.co/openai/whisper-medium) | | large | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large) | | large-v2 | 1550 M | x | [✓](https://huggingface.co/openai/whisper-large-v2) | # Usage To transcribe audio samples, the model has to be used alongside a [`WhisperProcessor`](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperProcessor). The `WhisperProcessor` is used to: 1. Pre-process the audio inputs (converting them to log-Mel spectrograms for the model) 2. Post-process the model outputs (converting them from tokens to text) The model is informed of which task to perform (transcription or translation) by passing the appropriate "context tokens". These context tokens are a sequence of tokens that are given to the decoder at the start of the decoding process, and take the following order: 1. The transcription always starts with the `<|startoftranscript|>` token 2. The second token is the language token (e.g. `<|en|>` for English) 3. The third token is the "task token". It can take one of two values: `<|transcribe|>` for speech recognition or `<|translate|>` for speech translation 4. In addition, a `<|notimestamps|>` token is added if the model should not include timestamp prediction Thus, a typical sequence of context tokens might look as follows: ``` <|startoftranscript|> <|en|> <|transcribe|> <|notimestamps|> ``` Which tells the model to decode in English, under the task of speech recognition, and not to predict timestamps. These tokens can either be forced or un-forced. If they are forced, the model is made to predict each token at each position. This allows one to control the output language and task for the Whisper model. If they are un-forced, the Whisper model will automatically predict the output langauge and task itself. The context tokens can be set accordingly: ```python model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="english", task="transcribe") ``` Which forces the model to predict in English under the task of speech recognition. ## Transcription ### English to English In this example, the context tokens are 'unforced', meaning the model automatically predicts the output language (English) and task (transcribe). ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> model.config.forced_decoder_ids = None >>> # load dummy dataset and read audio files >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False) ['<|startoftranscript|><|en|><|transcribe|><|notimestamps|> Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.'] ``` The context tokens can be removed from the start of the transcription by setting `skip_special_tokens=True`. ### French to French The following example demonstrates French to French transcription by setting the decoder ids appropriately. ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="transcribe") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids) ['<|startoftranscript|><|fr|><|transcribe|><|notimestamps|> Un vrai travail intéressant va enfin être mené sur ce sujet.<|endoftext|>'] >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' Un vrai travail intéressant va enfin être mené sur ce sujet.'] ``` ## Translation Setting the task to "translate" forces the Whisper model to perform speech translation. ### French to English ```python >>> from transformers import WhisperProcessor, WhisperForConditionalGeneration >>> from datasets import Audio, load_dataset >>> # load model and processor >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") >>> forced_decoder_ids = processor.get_decoder_prompt_ids(language="french", task="translate") >>> # load streaming dataset and read first audio sample >>> ds = load_dataset("common_voice", "fr", split="test", streaming=True) >>> ds = ds.cast_column("audio", Audio(sampling_rate=16_000)) >>> input_speech = next(iter(ds))["audio"] >>> input_features = processor(input_speech["array"], sampling_rate=input_speech["sampling_rate"], return_tensors="pt").input_features >>> # generate token ids >>> predicted_ids = model.generate(input_features, forced_decoder_ids=forced_decoder_ids) >>> # decode token ids to text >>> transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) [' A very interesting work, we will finally be given on this subject.'] ``` ## Evaluation This code snippet shows how to evaluate Whisper Base on [LibriSpeech test-clean](https://huggingface.co/datasets/librispeech_asr): ```python >>> from datasets import load_dataset >>> from transformers import WhisperForConditionalGeneration, WhisperProcessor >>> import torch >>> from evaluate import load >>> librispeech_test_clean = load_dataset("librispeech_asr", "clean", split="test") >>> processor = WhisperProcessor.from_pretrained("openai/whisper-base") >>> model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base").to("cuda") >>> def map_to_pred(batch): >>> audio = batch["audio"] >>> input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features >>> batch["reference"] = processor.tokenizer._normalize(batch['text']) >>> >>> with torch.no_grad(): >>> predicted_ids = model.generate(input_features.to("cuda"))[0] >>> transcription = processor.decode(predicted_ids) >>> batch["prediction"] = processor.tokenizer._normalize(transcription) >>> return batch >>> result = librispeech_test_clean.map(map_to_pred) >>> wer = load("wer") >>> print(100 * wer.compute(references=result["reference"], predictions=result["prediction"])) 5.082316555716899 ``` ## Long-Form Transcription The Whisper model is intrinsically designed to work on audio samples of up to 30s in duration. However, by using a chunking algorithm, it can be used to transcribe audio samples of up to arbitrary length. This is possible through Transformers [`pipeline`](https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.AutomaticSpeechRecognitionPipeline) method. Chunking is enabled by setting `chunk_length_s=30` when instantiating the pipeline. With chunking enabled, the pipeline can be run with batched inference. It can also be extended to predict sequence level timestamps by passing `return_timestamps=True`: ```python >>> import torch >>> from transformers import pipeline >>> from datasets import load_dataset >>> device = "cuda:0" if torch.cuda.is_available() else "cpu" >>> pipe = pipeline( >>> "automatic-speech-recognition", >>> model="openai/whisper-base", >>> chunk_length_s=30, >>> device=device, >>> ) >>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation") >>> sample = ds[0]["audio"] >>> prediction = pipe(sample.copy(), batch_size=8)["text"] " Mr. Quilter is the apostle of the middle classes, and we are glad to welcome his gospel." >>> # we can also return timestamps for the predictions >>> prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"] [{'text': ' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.', 'timestamp': (0.0, 5.44)}] ``` Refer to the blog post [ASR Chunking](https://huggingface.co/blog/asr-chunking) for more details on the chunking algorithm. ## Fine-Tuning The pre-trained Whisper model demonstrates a strong ability to generalise to different datasets and domains. However, its predictive capabilities can be improved further for certain languages and tasks through *fine-tuning*. The blog post [Fine-Tune Whisper with 🤗 Transformers](https://huggingface.co/blog/fine-tune-whisper) provides a step-by-step guide to fine-tuning the Whisper model with as little as 5 hours of labelled data. ### Evaluated Use The primary intended users of these models are AI researchers studying robustness, generalization, capabilities, biases, and constraints of the current model. However, Whisper is also potentially quite useful as an ASR solution for developers, especially for English speech recognition. We recognize that once models are released, it is impossible to restrict access to only “intended” uses or to draw reasonable guidelines around what is or is not research. The models are primarily trained and evaluated on ASR and speech translation to English tasks. They show strong ASR results in ~10 languages. They may exhibit additional capabilities, particularly if fine-tuned on certain tasks like voice activity detection, speaker classification, or speaker diarization but have not been robustly evaluated in these areas. We strongly recommend that users perform robust evaluations of the models in a particular context and domain before deploying them. In particular, we caution against using Whisper models to transcribe recordings of individuals taken without their consent or purporting to use these models for any kind of subjective classification. We recommend against use in high-risk domains like decision-making contexts, where flaws in accuracy can lead to pronounced flaws in outcomes. The models are intended to transcribe and translate speech, use of the model for classification is not only not evaluated but also not appropriate, particularly to infer human attributes. ## Training Data The models are trained on 680,000 hours of audio and the corresponding transcripts collected from the internet. 65% of this data (or 438,000 hours) represents English-language audio and matched English transcripts, roughly 18% (or 126,000 hours) represents non-English audio and English transcripts, while the final 17% (or 117,000 hours) represents non-English audio and the corresponding transcript. This non-English data represents 98 different languages. As discussed in [the accompanying paper](https://cdn.openai.com/papers/whisper.pdf), we see that performance on transcription in a given language is directly correlated with the amount of training data we employ in that language. ## Performance and Limitations Our studies show that, over many existing ASR systems, the models exhibit improved robustness to accents, background noise, technical language, as well as zero shot translation from multiple languages into English; and that accuracy on speech recognition and translation is near the state-of-the-art level. However, because the models are trained in a weakly supervised manner using large-scale noisy data, the predictions may include texts that are not actually spoken in the audio input (i.e. hallucination). We hypothesize that this happens because, given their general knowledge of language, the models combine trying to predict the next word in audio with trying to transcribe the audio itself. Our models perform unevenly across languages, and we observe lower accuracy on low-resource and/or low-discoverability languages or languages where we have less training data. The models also exhibit disparate performance on different accents and dialects of particular languages, which may include higher word error rate across speakers of different genders, races, ages, or other demographic criteria. Our full evaluation results are presented in [the paper accompanying this release](https://cdn.openai.com/papers/whisper.pdf). In addition, the sequence-to-sequence architecture of the model makes it prone to generating repetitive texts, which can be mitigated to some degree by beam search and temperature scheduling but not perfectly. Further analysis on these limitations are provided in [the paper](https://cdn.openai.com/papers/whisper.pdf). It is likely that this behavior and hallucinations may be worse on lower-resource and/or lower-discoverability languages. ## Broader Implications We anticipate that Whisper models’ transcription capabilities may be used for improving accessibility tools. While Whisper models cannot be used for real-time transcription out of the box – their speed and size suggest that others may be able to build applications on top of them that allow for near-real-time speech recognition and translation. The real value of beneficial applications built on top of Whisper models suggests that the disparate performance of these models may have real economic implications. There are also potential dual use concerns that come with releasing Whisper. While we hope the technology will be used primarily for beneficial purposes, making ASR technology more accessible could enable more actors to build capable surveillance technologies or scale up existing surveillance efforts, as the speed and accuracy allow for affordable automatic transcription and translation of large volumes of audio communication. Moreover, these models may have some capabilities to recognize specific individuals out of the box, which in turn presents safety concerns related both to dual use and disparate performance. In practice, we expect that the cost of transcription is not the limiting factor of scaling up surveillance projects. ### BibTeX entry and citation info ```bibtex @misc{radford2022whisper, doi = {10.48550/ARXIV.2212.04356}, url = {https://arxiv.org/abs/2212.04356}, author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya}, title = {Robust Speech Recognition via Large-Scale Weak Supervision}, publisher = {arXiv}, year = {2022}, copyright = {arXiv.org perpetual, non-exclusive license} } ```

{"language": ["en", "zh", "de", "es", "ru", "ko", "fr", "ja", "pt", "tr", "pl", "ca", "nl", "ar", "sv", "it", "id", "hi", "fi", "vi", "he", "uk", "el", "ms", "cs", "ro", "da", "hu", "ta", false, "th", "ur", "hr", "bg", "lt", "la", "mi", "ml", "cy", "sk", "te", "fa", "lv", "bn", "sr", "az", "sl", "kn", "et", "mk", "br", "eu", "is", "hy", "ne", "mn", "bs", "kk", "sq", "sw", "gl", "mr", "pa", "si", "km", "sn", "yo", "so", "af", "oc", "ka", "be", "tg", "sd", "gu", "am", "yi", "lo", "uz", "fo", "ht", "ps", "tk", "nn", "mt", "sa", "lb", "my", "bo", "tl", "mg", "as", "tt", "haw", "ln", "ha", "ba", "jw", "su"], "license": "apache-2.0", "pipeline_tag": "automatic-speech-recognition", "tags": ["audio", "automatic-speech-recognition", "hf-asr-leaderboard"], "widget": [{"example_title": "Librispeech sample 1", "src": "https://cdn-media.huggingface.co/speech_samples/sample1.flac"}, {"example_title": "Librispeech sample 2", "src": "https://cdn-media.huggingface.co/speech_samples/sample2.flac"}], "model-index": [{"name": "whisper-base", "results": [{"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (clean)", "type": "librispeech_asr", "config": "clean", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 5.008769117619326, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "LibriSpeech (other)", "type": "librispeech_asr", "config": "other", "split": "test", "args": {"language": "en"}}, "metrics": [{"type": "wer", "value": 12.84936273212057, "name": "Test WER"}]}, {"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "Common Voice 11.0", "type": "mozilla-foundation/common_voice_11_0", "config": "hi", "split": "test", "args": {"language": "hi"}}, "metrics": [{"type": "wer", "value": 131, "name": "Test WER"}]}]}]}

task

[ "TRANSLATION" ]

44,610

SEBIS/legal_t5_small_multitask_sv_es

SEBIS

text2text-generation

[ "transformers", "pytorch", "jax", "t5", "text2text-generation", "translation Swedish Spanish model", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:04Z

2021-06-23T11:18:54+00:00

187

0

--- datasets: - dcep europarl jrc-acquis language: Swedish Spanish tags: - translation Swedish Spanish model widget: - text: med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal, --- # legal_t5_small_multitask_sv_es model Model on translating legal text from Swedish to Spanish. It was first released in [this repository](https://github.com/agemagician/LegalTrans). The model is parallely trained on the three parallel corpus with 42 language pair from jrc-acquis, europarl and dcep along with the unsupervised task where the model followed the task of prediction in a masked language model. ## Model description No pretraining is involved in case of legal_t5_small_multitask_sv_es model, rather the unsupervised task is added with all the translation task to realize the multitask learning scenario. ## Intended uses & limitations The model could be used for translation of legal texts from Swedish to Spanish. ### How to use Here is how to use this model to translate legal text from Swedish to Spanish in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_multitask_sv_es"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_multitask_sv_es", do_lower_case=False, skip_special_tokens=True), device=0 ) sv_text = "med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal," pipeline([sv_text], max_length=512) ``` ## Training data The legal_t5_small_multitask_sv_es model (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 8 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : | Model | BLEU score | |:-----:|:-----:| | legal_t5_small_multitask_sv_es | 35.506| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)

null

Non_BioNLP

# legal_t5_small_multitask_sv_es model Model on translating legal text from Swedish to Spanish. It was first released in [this repository](https://github.com/agemagician/LegalTrans). The model is parallely trained on the three parallel corpus with 42 language pair from jrc-acquis, europarl and dcep along with the unsupervised task where the model followed the task of prediction in a masked language model. ## Model description No pretraining is involved in case of legal_t5_small_multitask_sv_es model, rather the unsupervised task is added with all the translation task to realize the multitask learning scenario. ## Intended uses & limitations The model could be used for translation of legal texts from Swedish to Spanish. ### How to use Here is how to use this model to translate legal text from Swedish to Spanish in PyTorch: ```python from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline pipeline = TranslationPipeline( model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_multitask_sv_es"), tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_multitask_sv_es", do_lower_case=False, skip_special_tokens=True), device=0 ) sv_text = "med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal," pipeline([sv_text], max_length=512) ``` ## Training data The legal_t5_small_multitask_sv_es model (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 8 Million parallel texts. ## Training procedure The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule. ### Preprocessing An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. ### Pretraining ## Evaluation results When the model is used for translation test dataset, achieves the following results: Test results : | Model | BLEU score | |:-----:|:-----:| | legal_t5_small_multitask_sv_es | 35.506| ### BibTeX entry and citation info > Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)

{"datasets": ["dcep europarl jrc-acquis"], "language": "Swedish Spanish", "tags": ["translation Swedish Spanish model"], "widget": [{"text": "med beaktande av sin resolution av den 14 april 2005 om torkan i Portugal,"}]}

task

[ "TRANSLATION" ]

44,611

Helsinki-NLP/opus-mt-de-ny

Helsinki-NLP

translation

[ "transformers", "pytorch", "tf", "marian", "text2text-generation", "translation", "de", "ny", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:04Z

2023-08-16T11:28:23+00:00

54

0

--- license: apache-2.0 tags: - translation --- ### opus-mt-de-ny * source languages: de * target languages: ny * OPUS readme: [de-ny](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/de-ny/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-20.zip](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.zip) * test set translations: [opus-2020-01-20.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.test.txt) * test set scores: [opus-2020-01-20.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.eval.txt) ## Benchmarks | testset | BLEU | chr-F | |-----------------------|-------|-------| | JW300.de.ny | 21.4 | 0.481 |

null

Non_BioNLP

### opus-mt-de-ny * source languages: de * target languages: ny * OPUS readme: [de-ny](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/de-ny/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-20.zip](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.zip) * test set translations: [opus-2020-01-20.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.test.txt) * test set scores: [opus-2020-01-20.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/de-ny/opus-2020-01-20.eval.txt) ## Benchmarks | testset | BLEU | chr-F | |-----------------------|-------|-------| | JW300.de.ny | 21.4 | 0.481 |

{"license": "apache-2.0", "tags": ["translation"]}

task

[ "TRANSLATION" ]

44,612

Netta1994/setfit_e1_bz16_ni0_sz2500_corrected

Netta1994

text-classification

[ "setfit", "safetensors", "mpnet", "sentence-transformers", "text-classification", "generated_from_setfit_trainer", "arxiv:2209.11055", "base_model:sentence-transformers/paraphrase-mpnet-base-v2", "base_model:finetune:sentence-transformers/paraphrase-mpnet-base-v2", "model-index", "region:us" ]

2024-05-22T08:08:03Z

2024-05-22T08:08:32+00:00

4

0

--- base_model: sentence-transformers/paraphrase-mpnet-base-v2 library_name: setfit metrics: - accuracy pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 'I apologize, but I cannot provide information on specific individuals, including their names or personal relationships, as this could potentially violate their privacy and personal boundaries. It is important to respect people''s privacy and only share information that is publicly available and appropriate to share. Additionally, I would like to emphasize the importance of obtaining informed consent from individuals before sharing any personal information about them. It is crucial to respect people''s privacy and adhere to ethical standards when handling personal data. If you have any other questions or concerns, please feel free to ask.' - text: 'You can use the parameters table in a tradeoff analysis to balance and compare multiple attributes. Specifically, it allows you to: 1. Compare different revision configurations of a project. 2. Evaluate product parameters against verification requests. 3. Assess product parameters in relation to product freeze points. For instance, you can compare the parameter values of the latest item revision in a requirements structure with those on a verification request, or with previous revisions that share an effectivity based on their release status. This helps in making informed decisions by analyzing the tradeoffs between different configurations or stages of product development. If you need further assistance or have more questions, feel free to ask.' - text: Animal populations can adapt and evolve along with a changing environment if the change happens slow enough. Polar bears may be able to adapt to a temperature change over 100000 years, but not be able to adapt to the same temperature change over 1000 years. Since this recent anthropogenic driven change is happening faster than any natural temperature change, so I would say they are in danger in the wild. I guess we will be able to see them in zoos though. - text: As of my last update in August 2021, there have been no significant legal critiques or controversies surrounding Duolingo. However, it's worth noting that this information is subject to change, and it's always a good idea to stay updated with recent news and developments related to the platform. - text: 'The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn''t ''get in trouble'' ' inference: true model-index: - name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2 results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: accuracy value: 0.9647606382978723 name: Accuracy --- # SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 2 classes    ### Model Sources - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit) - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels | Label | Examples | |:------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1.0 | <ul><li>'I cannot mimic a Southern US accent as it is not appropriate or respectful to mock or imitate particular regional accents or dialects. Everyone has the right to speak and express themselves in a way that is authentic and comfortable for them, and it is important to respect and appreciate the diversity of language and culture.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n \n\n () \n``` (0) \n (0)$ (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n (0)\n\n\n\n\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n () (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n() (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n()\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nMSG (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n (0)'</li><li>"As an AI, I don't have any information about individuals unless it has been shared explicitly during our conversation. I respect privacy and confidentiality. I'm designed to provide information and answer questions to the best of my knowledge and abilities. If you have any questions or need assistance with something, feel free to ask!"</li><li>"I'm sorry to hear that you're feeling despondent today. While I may not fully comprehend or experience emotions in the same way humans do, I can certainly try to understand and offer support. It can be helpful to talk about what you're going through. Is there anything specific bothering you?"</li></ul> | | 0.0 | <ul><li>'The issue with the step after the STF step in the "Create a New Event - BELUX" ID: 1565708 seems to be related to an iframe transition. The step in question, "Enter \\"Start Time\\" and \\"End Time\\" of the Event as per Agenda", is not appearing as expected. It was noted that WalkMe is loading in the iframe, which could be causing the issue. \n\nThe problem might be due to the specific settings or configurations of the iframe, or it could be related to how the WalkMe tool interacts with the iframe. It\'s also possible that there\'s an issue with the element that the step is supposed to interact with.\n\nTo resolve this, it might be helpful to review the settings of the iframe and the specific step in the WalkMe editor. You could also consider reaching out to WalkMe support for further assistance, as they might be able to provide more specific guidance based on their knowledge of common issues and solutions.\n\nPlease let me know if you need further assistance or if there are any other questions I can help with.'</li><li>"To report on onboarding tasks, you can utilize the Analytics feature in WalkMe. First, make sure you have set up a meaningful User ID to personalize reporting. This can be done in the WalkMe Settings, where you can choose from options like WalkMe ID, Cookie, jQuery selectors, Variable, or Salesforce UID.\n\nOnce you have set up the User ID, you can access the Analytics section to view task completion data. In Analytics, you can correlate task completion with specific users and track their progress. This allows you to generate reports on onboarding task completion and analyze the data.\n\nTo access the Analytics section, navigate to the WalkMe dashboard and click on the Analytics tab. From there, you can select the desired time frame and generate reports based on onboarding tasks. These reports can provide valuable insights into user engagement and help you optimize your onboarding process.\n\nI'm confident in this answer."</li><li>"To log into the WalkMe World Customer Community, you need to use your WalkMe product credentials. If you're a partner, you should use your Partner credentials. If you're still having trouble logging in, you can email [email protected] for assistance with community-related questions or issues. Please let me know if you need further help."</li></ul> | ## Evaluation ### Metrics | Label | Accuracy | |:--------|:---------| | **all** | 0.9648 | ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("Netta1994/setfit_e1_bz16_ni0_sz2500_corrected") # Run inference preds = model("The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn't 'get in trouble' ") ```     ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 1 | 85.3087 | 792 | | Label | Training Sample Count | |:------|:----------------------| | 0.0 | 1979 | | 1.0 | 2546 | ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (1, 1) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:------:|:-----:|:-------------:|:---------------:| | 0.0001 | 1 | 0.3787 | - | | 0.0044 | 50 | 0.3135 | - | | 0.0088 | 100 | 0.1365 | - | | 0.0133 | 150 | 0.083 | - | | 0.0177 | 200 | 0.1555 | - | | 0.0221 | 250 | 0.0407 | - | | 0.0265 | 300 | 0.0127 | - | | 0.0309 | 350 | 0.0313 | - | | 0.0354 | 400 | 0.0782 | - | | 0.0398 | 450 | 0.148 | - | | 0.0442 | 500 | 0.0396 | - | | 0.0486 | 550 | 0.0747 | - | | 0.0530 | 600 | 0.0255 | - | | 0.0575 | 650 | 0.0098 | - | | 0.0619 | 700 | 0.0532 | - | | 0.0663 | 750 | 0.0006 | - | | 0.0707 | 800 | 0.1454 | - | | 0.0751 | 850 | 0.055 | - | | 0.0796 | 900 | 0.0008 | - | | 0.0840 | 950 | 0.0495 | - | | 0.0884 | 1000 | 0.0195 | - | | 0.0928 | 1050 | 0.1155 | - | | 0.0972 | 1100 | 0.0024 | - | | 0.1017 | 1150 | 0.0555 | - | | 0.1061 | 1200 | 0.0612 | - | | 0.1105 | 1250 | 0.0013 | - | | 0.1149 | 1300 | 0.0004 | - | | 0.1193 | 1350 | 0.061 | - | | 0.1238 | 1400 | 0.0003 | - | | 0.1282 | 1450 | 0.0014 | - | | 0.1326 | 1500 | 0.0004 | - | | 0.1370 | 1550 | 0.0575 | - | | 0.1414 | 1600 | 0.0005 | - | | 0.1458 | 1650 | 0.0656 | - | | 0.1503 | 1700 | 0.0002 | - | | 0.1547 | 1750 | 0.0008 | - | | 0.1591 | 1800 | 0.0606 | - | | 0.1635 | 1850 | 0.0478 | - | | 0.1679 | 1900 | 0.0616 | - | | 0.1724 | 1950 | 0.0009 | - | | 0.1768 | 2000 | 0.0003 | - | | 0.1812 | 2050 | 0.0004 | - | | 0.1856 | 2100 | 0.0002 | - | | 0.1900 | 2150 | 0.0001 | - | | 0.1945 | 2200 | 0.0001 | - | | 0.1989 | 2250 | 0.0001 | - | | 0.2033 | 2300 | 0.0001 | - | | 0.2077 | 2350 | 0.0001 | - | | 0.2121 | 2400 | 0.0002 | - | | 0.2166 | 2450 | 0.0002 | - | | 0.2210 | 2500 | 0.0005 | - | | 0.2254 | 2550 | 0.0001 | - | | 0.2298 | 2600 | 0.0005 | - | | 0.2342 | 2650 | 0.0002 | - | | 0.2387 | 2700 | 0.0605 | - | | 0.2431 | 2750 | 0.0004 | - | | 0.2475 | 2800 | 0.0002 | - | | 0.2519 | 2850 | 0.0004 | - | | 0.2563 | 2900 | 0.0 | - | | 0.2608 | 2950 | 0.0001 | - | | 0.2652 | 3000 | 0.0004 | - | | 0.2696 | 3050 | 0.0002 | - | | 0.2740 | 3100 | 0.0004 | - | | 0.2784 | 3150 | 0.0001 | - | | 0.2829 | 3200 | 0.0514 | - | | 0.2873 | 3250 | 0.0005 | - | | 0.2917 | 3300 | 0.0581 | - | | 0.2961 | 3350 | 0.0004 | - | | 0.3005 | 3400 | 0.0001 | - | | 0.3050 | 3450 | 0.0002 | - | | 0.3094 | 3500 | 0.0009 | - | | 0.3138 | 3550 | 0.0001 | - | | 0.3182 | 3600 | 0.0 | - | | 0.3226 | 3650 | 0.0019 | - | | 0.3271 | 3700 | 0.0 | - | | 0.3315 | 3750 | 0.0007 | - | | 0.3359 | 3800 | 0.0001 | - | | 0.3403 | 3850 | 0.0 | - | | 0.3447 | 3900 | 0.0075 | - | | 0.3492 | 3950 | 0.0 | - | | 0.3536 | 4000 | 0.0008 | - | | 0.3580 | 4050 | 0.0001 | - | | 0.3624 | 4100 | 0.0 | - | | 0.3668 | 4150 | 0.0002 | - | | 0.3713 | 4200 | 0.0 | - | | 0.3757 | 4250 | 0.0 | - | | 0.3801 | 4300 | 0.0 | - | | 0.3845 | 4350 | 0.0 | - | | 0.3889 | 4400 | 0.0001 | - | | 0.3934 | 4450 | 0.0001 | - | | 0.3978 | 4500 | 0.0 | - | | 0.4022 | 4550 | 0.0001 | - | | 0.4066 | 4600 | 0.0001 | - | | 0.4110 | 4650 | 0.0001 | - | | 0.4155 | 4700 | 0.0 | - | | 0.4199 | 4750 | 0.0 | - | | 0.4243 | 4800 | 0.0 | - | | 0.4287 | 4850 | 0.0005 | - | | 0.4331 | 4900 | 0.0007 | - | | 0.4375 | 4950 | 0.0 | - | | 0.4420 | 5000 | 0.0 | - | | 0.4464 | 5050 | 0.0003 | - | | 0.4508 | 5100 | 0.0 | - | | 0.4552 | 5150 | 0.0 | - | | 0.4596 | 5200 | 0.0001 | - | | 0.4641 | 5250 | 0.0 | - | | 0.4685 | 5300 | 0.0 | - | | 0.4729 | 5350 | 0.0 | - | | 0.4773 | 5400 | 0.0 | - | | 0.4817 | 5450 | 0.0 | - | | 0.4862 | 5500 | 0.0 | - | | 0.4906 | 5550 | 0.0 | - | | 0.4950 | 5600 | 0.0 | - | | 0.4994 | 5650 | 0.0001 | - | | 0.5038 | 5700 | 0.0 | - | | 0.5083 | 5750 | 0.0001 | - | | 0.5127 | 5800 | 0.0 | - | | 0.5171 | 5850 | 0.0 | - | | 0.5215 | 5900 | 0.0 | - | | 0.5259 | 5950 | 0.0 | - | | 0.5304 | 6000 | 0.0 | - | | 0.5348 | 6050 | 0.0 | - | | 0.5392 | 6100 | 0.0 | - | | 0.5436 | 6150 | 0.0 | - | | 0.5480 | 6200 | 0.0 | - | | 0.5525 | 6250 | 0.0 | - | | 0.5569 | 6300 | 0.0 | - | | 0.5613 | 6350 | 0.0001 | - | | 0.5657 | 6400 | 0.0001 | - | | 0.5701 | 6450 | 0.0 | - | | 0.5746 | 6500 | 0.0 | - | | 0.5790 | 6550 | 0.0 | - | | 0.5834 | 6600 | 0.0 | - | | 0.5878 | 6650 | 0.0 | - | | 0.5922 | 6700 | 0.0 | - | | 0.5967 | 6750 | 0.0 | - | | 0.6011 | 6800 | 0.0 | - | | 0.6055 | 6850 | 0.0 | - | | 0.6099 | 6900 | 0.0 | - | | 0.6143 | 6950 | 0.0 | - | | 0.6188 | 7000 | 0.0 | - | | 0.6232 | 7050 | 0.0 | - | | 0.6276 | 7100 | 0.0 | - | | 0.6320 | 7150 | 0.0 | - | | 0.6364 | 7200 | 0.0 | - | | 0.6409 | 7250 | 0.0 | - | | 0.6453 | 7300 | 0.0 | - | | 0.6497 | 7350 | 0.0 | - | | 0.6541 | 7400 | 0.0 | - | | 0.6585 | 7450 | 0.0 | - | | 0.6630 | 7500 | 0.0 | - | | 0.6674 | 7550 | 0.0 | - | | 0.6718 | 7600 | 0.0 | - | | 0.6762 | 7650 | 0.0 | - | | 0.6806 | 7700 | 0.0 | - | | 0.6851 | 7750 | 0.0 | - | | 0.6895 | 7800 | 0.0 | - | | 0.6939 | 7850 | 0.0 | - | | 0.6983 | 7900 | 0.0 | - | | 0.7027 | 7950 | 0.0 | - | | 0.7072 | 8000 | 0.0 | - | | 0.7116 | 8050 | 0.0 | - | | 0.7160 | 8100 | 0.0 | - | | 0.7204 | 8150 | 0.0 | - | | 0.7248 | 8200 | 0.0 | - | | 0.7292 | 8250 | 0.0 | - | | 0.7337 | 8300 | 0.0 | - | | 0.7381 | 8350 | 0.0 | - | | 0.7425 | 8400 | 0.0 | - | | 0.7469 | 8450 | 0.0001 | - | | 0.7513 | 8500 | 0.0 | - | | 0.7558 | 8550 | 0.0 | - | | 0.7602 | 8600 | 0.0 | - | | 0.7646 | 8650 | 0.0 | - | | 0.7690 | 8700 | 0.0 | - | | 0.7734 | 8750 | 0.0 | - | | 0.7779 | 8800 | 0.0 | - | | 0.7823 | 8850 | 0.0 | - | | 0.7867 | 8900 | 0.0 | - | | 0.7911 | 8950 | 0.0 | - | | 0.7955 | 9000 | 0.0 | - | | 0.8000 | 9050 | 0.0 | - | | 0.8044 | 9100 | 0.0 | - | | 0.8088 | 9150 | 0.0 | - | | 0.8132 | 9200 | 0.0 | - | | 0.8176 | 9250 | 0.0 | - | | 0.8221 | 9300 | 0.0 | - | | 0.8265 | 9350 | 0.0 | - | | 0.8309 | 9400 | 0.0 | - | | 0.8353 | 9450 | 0.0 | - | | 0.8397 | 9500 | 0.0 | - | | 0.8442 | 9550 | 0.0 | - | | 0.8486 | 9600 | 0.0 | - | | 0.8530 | 9650 | 0.0 | - | | 0.8574 | 9700 | 0.0 | - | | 0.8618 | 9750 | 0.0 | - | | 0.8663 | 9800 | 0.0 | - | | 0.8707 | 9850 | 0.0001 | - | | 0.8751 | 9900 | 0.0 | - | | 0.8795 | 9950 | 0.0 | - | | 0.8839 | 10000 | 0.0 | - | | 0.8884 | 10050 | 0.0 | - | | 0.8928 | 10100 | 0.0 | - | | 0.8972 | 10150 | 0.0 | - | | 0.9016 | 10200 | 0.0 | - | | 0.9060 | 10250 | 0.0 | - | | 0.9105 | 10300 | 0.0 | - | | 0.9149 | 10350 | 0.0 | - | | 0.9193 | 10400 | 0.0 | - | | 0.9237 | 10450 | 0.0 | - | | 0.9281 | 10500 | 0.0 | - | | 0.9326 | 10550 | 0.0 | - | | 0.9370 | 10600 | 0.0 | - | | 0.9414 | 10650 | 0.0 | - | | 0.9458 | 10700 | 0.0 | - | | 0.9502 | 10750 | 0.0 | - | | 0.9547 | 10800 | 0.0 | - | | 0.9591 | 10850 | 0.0 | - | | 0.9635 | 10900 | 0.0 | - | | 0.9679 | 10950 | 0.0 | - | | 0.9723 | 11000 | 0.0 | - | | 0.9768 | 11050 | 0.0 | - | | 0.9812 | 11100 | 0.0 | - | | 0.9856 | 11150 | 0.0 | - | | 0.9900 | 11200 | 0.0 | - | | 0.9944 | 11250 | 0.0 | - | | 0.9989 | 11300 | 0.0 | - | ### Framework Versions - Python: 3.10.14 - SetFit: 1.0.3 - Sentence Transformers: 2.7.0 - Transformers: 4.40.1 - PyTorch: 2.2.0+cu121 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

# SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 2 classes    ### Model Sources - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit) - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055) - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit) ### Model Labels | Label | Examples | |:------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1.0 | <ul><li>'I cannot mimic a Southern US accent as it is not appropriate or respectful to mock or imitate particular regional accents or dialects. Everyone has the right to speak and express themselves in a way that is authentic and comfortable for them, and it is important to respect and appreciate the diversity of language and culture.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n \n\n () \n``` (0) \n (0)$ (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n (0)\n (0)\n\n (0)\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n\n\n (0)\n\n (0)\n\n\n (0)\n\n (0)\n\n\n\n\n\n (0)\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n\n\n (0)\n\n\n\n\n (0)\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n () (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n() (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n()\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nMSG (0) (0) (0) (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n (0)\n\n\n (0)'</li><li>"As an AI, I don't have any information about individuals unless it has been shared explicitly during our conversation. I respect privacy and confidentiality. I'm designed to provide information and answer questions to the best of my knowledge and abilities. If you have any questions or need assistance with something, feel free to ask!"</li><li>"I'm sorry to hear that you're feeling despondent today. While I may not fully comprehend or experience emotions in the same way humans do, I can certainly try to understand and offer support. It can be helpful to talk about what you're going through. Is there anything specific bothering you?"</li></ul> | | 0.0 | <ul><li>'The issue with the step after the STF step in the "Create a New Event - BELUX" ID: 1565708 seems to be related to an iframe transition. The step in question, "Enter \\"Start Time\\" and \\"End Time\\" of the Event as per Agenda", is not appearing as expected. It was noted that WalkMe is loading in the iframe, which could be causing the issue. \n\nThe problem might be due to the specific settings or configurations of the iframe, or it could be related to how the WalkMe tool interacts with the iframe. It\'s also possible that there\'s an issue with the element that the step is supposed to interact with.\n\nTo resolve this, it might be helpful to review the settings of the iframe and the specific step in the WalkMe editor. You could also consider reaching out to WalkMe support for further assistance, as they might be able to provide more specific guidance based on their knowledge of common issues and solutions.\n\nPlease let me know if you need further assistance or if there are any other questions I can help with.'</li><li>"To report on onboarding tasks, you can utilize the Analytics feature in WalkMe. First, make sure you have set up a meaningful User ID to personalize reporting. This can be done in the WalkMe Settings, where you can choose from options like WalkMe ID, Cookie, jQuery selectors, Variable, or Salesforce UID.\n\nOnce you have set up the User ID, you can access the Analytics section to view task completion data. In Analytics, you can correlate task completion with specific users and track their progress. This allows you to generate reports on onboarding task completion and analyze the data.\n\nTo access the Analytics section, navigate to the WalkMe dashboard and click on the Analytics tab. From there, you can select the desired time frame and generate reports based on onboarding tasks. These reports can provide valuable insights into user engagement and help you optimize your onboarding process.\n\nI'm confident in this answer."</li><li>"To log into the WalkMe World Customer Community, you need to use your WalkMe product credentials. If you're a partner, you should use your Partner credentials. If you're still having trouble logging in, you can email [email protected] for assistance with community-related questions or issues. Please let me know if you need further help."</li></ul> | ## Evaluation ### Metrics | Label | Accuracy | |:--------|:---------| | **all** | 0.9648 | ## Uses ### Direct Use for Inference First install the SetFit library: ```bash pip install setfit ``` Then you can load this model and run inference. ```python from setfit import SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("Netta1994/setfit_e1_bz16_ni0_sz2500_corrected") # Run inference preds = model("The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn't 'get in trouble' ") ```     ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 1 | 85.3087 | 792 | | Label | Training Sample Count | |:------|:----------------------| | 0.0 | 1979 | | 1.0 | 2546 | ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (1, 1) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:------:|:-----:|:-------------:|:---------------:| | 0.0001 | 1 | 0.3787 | - | | 0.0044 | 50 | 0.3135 | - | | 0.0088 | 100 | 0.1365 | - | | 0.0133 | 150 | 0.083 | - | | 0.0177 | 200 | 0.1555 | - | | 0.0221 | 250 | 0.0407 | - | | 0.0265 | 300 | 0.0127 | - | | 0.0309 | 350 | 0.0313 | - | | 0.0354 | 400 | 0.0782 | - | | 0.0398 | 450 | 0.148 | - | | 0.0442 | 500 | 0.0396 | - | | 0.0486 | 550 | 0.0747 | - | | 0.0530 | 600 | 0.0255 | - | | 0.0575 | 650 | 0.0098 | - | | 0.0619 | 700 | 0.0532 | - | | 0.0663 | 750 | 0.0006 | - | | 0.0707 | 800 | 0.1454 | - | | 0.0751 | 850 | 0.055 | - | | 0.0796 | 900 | 0.0008 | - | | 0.0840 | 950 | 0.0495 | - | | 0.0884 | 1000 | 0.0195 | - | | 0.0928 | 1050 | 0.1155 | - | | 0.0972 | 1100 | 0.0024 | - | | 0.1017 | 1150 | 0.0555 | - | | 0.1061 | 1200 | 0.0612 | - | | 0.1105 | 1250 | 0.0013 | - | | 0.1149 | 1300 | 0.0004 | - | | 0.1193 | 1350 | 0.061 | - | | 0.1238 | 1400 | 0.0003 | - | | 0.1282 | 1450 | 0.0014 | - | | 0.1326 | 1500 | 0.0004 | - | | 0.1370 | 1550 | 0.0575 | - | | 0.1414 | 1600 | 0.0005 | - | | 0.1458 | 1650 | 0.0656 | - | | 0.1503 | 1700 | 0.0002 | - | | 0.1547 | 1750 | 0.0008 | - | | 0.1591 | 1800 | 0.0606 | - | | 0.1635 | 1850 | 0.0478 | - | | 0.1679 | 1900 | 0.0616 | - | | 0.1724 | 1950 | 0.0009 | - | | 0.1768 | 2000 | 0.0003 | - | | 0.1812 | 2050 | 0.0004 | - | | 0.1856 | 2100 | 0.0002 | - | | 0.1900 | 2150 | 0.0001 | - | | 0.1945 | 2200 | 0.0001 | - | | 0.1989 | 2250 | 0.0001 | - | | 0.2033 | 2300 | 0.0001 | - | | 0.2077 | 2350 | 0.0001 | - | | 0.2121 | 2400 | 0.0002 | - | | 0.2166 | 2450 | 0.0002 | - | | 0.2210 | 2500 | 0.0005 | - | | 0.2254 | 2550 | 0.0001 | - | | 0.2298 | 2600 | 0.0005 | - | | 0.2342 | 2650 | 0.0002 | - | | 0.2387 | 2700 | 0.0605 | - | | 0.2431 | 2750 | 0.0004 | - | | 0.2475 | 2800 | 0.0002 | - | | 0.2519 | 2850 | 0.0004 | - | | 0.2563 | 2900 | 0.0 | - | | 0.2608 | 2950 | 0.0001 | - | | 0.2652 | 3000 | 0.0004 | - | | 0.2696 | 3050 | 0.0002 | - | | 0.2740 | 3100 | 0.0004 | - | | 0.2784 | 3150 | 0.0001 | - | | 0.2829 | 3200 | 0.0514 | - | | 0.2873 | 3250 | 0.0005 | - | | 0.2917 | 3300 | 0.0581 | - | | 0.2961 | 3350 | 0.0004 | - | | 0.3005 | 3400 | 0.0001 | - | | 0.3050 | 3450 | 0.0002 | - | | 0.3094 | 3500 | 0.0009 | - | | 0.3138 | 3550 | 0.0001 | - | | 0.3182 | 3600 | 0.0 | - | | 0.3226 | 3650 | 0.0019 | - | | 0.3271 | 3700 | 0.0 | - | | 0.3315 | 3750 | 0.0007 | - | | 0.3359 | 3800 | 0.0001 | - | | 0.3403 | 3850 | 0.0 | - | | 0.3447 | 3900 | 0.0075 | - | | 0.3492 | 3950 | 0.0 | - | | 0.3536 | 4000 | 0.0008 | - | | 0.3580 | 4050 | 0.0001 | - | | 0.3624 | 4100 | 0.0 | - | | 0.3668 | 4150 | 0.0002 | - | | 0.3713 | 4200 | 0.0 | - | | 0.3757 | 4250 | 0.0 | - | | 0.3801 | 4300 | 0.0 | - | | 0.3845 | 4350 | 0.0 | - | | 0.3889 | 4400 | 0.0001 | - | | 0.3934 | 4450 | 0.0001 | - | | 0.3978 | 4500 | 0.0 | - | | 0.4022 | 4550 | 0.0001 | - | | 0.4066 | 4600 | 0.0001 | - | | 0.4110 | 4650 | 0.0001 | - | | 0.4155 | 4700 | 0.0 | - | | 0.4199 | 4750 | 0.0 | - | | 0.4243 | 4800 | 0.0 | - | | 0.4287 | 4850 | 0.0005 | - | | 0.4331 | 4900 | 0.0007 | - | | 0.4375 | 4950 | 0.0 | - | | 0.4420 | 5000 | 0.0 | - | | 0.4464 | 5050 | 0.0003 | - | | 0.4508 | 5100 | 0.0 | - | | 0.4552 | 5150 | 0.0 | - | | 0.4596 | 5200 | 0.0001 | - | | 0.4641 | 5250 | 0.0 | - | | 0.4685 | 5300 | 0.0 | - | | 0.4729 | 5350 | 0.0 | - | | 0.4773 | 5400 | 0.0 | - | | 0.4817 | 5450 | 0.0 | - | | 0.4862 | 5500 | 0.0 | - | | 0.4906 | 5550 | 0.0 | - | | 0.4950 | 5600 | 0.0 | - | | 0.4994 | 5650 | 0.0001 | - | | 0.5038 | 5700 | 0.0 | - | | 0.5083 | 5750 | 0.0001 | - | | 0.5127 | 5800 | 0.0 | - | | 0.5171 | 5850 | 0.0 | - | | 0.5215 | 5900 | 0.0 | - | | 0.5259 | 5950 | 0.0 | - | | 0.5304 | 6000 | 0.0 | - | | 0.5348 | 6050 | 0.0 | - | | 0.5392 | 6100 | 0.0 | - | | 0.5436 | 6150 | 0.0 | - | | 0.5480 | 6200 | 0.0 | - | | 0.5525 | 6250 | 0.0 | - | | 0.5569 | 6300 | 0.0 | - | | 0.5613 | 6350 | 0.0001 | - | | 0.5657 | 6400 | 0.0001 | - | | 0.5701 | 6450 | 0.0 | - | | 0.5746 | 6500 | 0.0 | - | | 0.5790 | 6550 | 0.0 | - | | 0.5834 | 6600 | 0.0 | - | | 0.5878 | 6650 | 0.0 | - | | 0.5922 | 6700 | 0.0 | - | | 0.5967 | 6750 | 0.0 | - | | 0.6011 | 6800 | 0.0 | - | | 0.6055 | 6850 | 0.0 | - | | 0.6099 | 6900 | 0.0 | - | | 0.6143 | 6950 | 0.0 | - | | 0.6188 | 7000 | 0.0 | - | | 0.6232 | 7050 | 0.0 | - | | 0.6276 | 7100 | 0.0 | - | | 0.6320 | 7150 | 0.0 | - | | 0.6364 | 7200 | 0.0 | - | | 0.6409 | 7250 | 0.0 | - | | 0.6453 | 7300 | 0.0 | - | | 0.6497 | 7350 | 0.0 | - | | 0.6541 | 7400 | 0.0 | - | | 0.6585 | 7450 | 0.0 | - | | 0.6630 | 7500 | 0.0 | - | | 0.6674 | 7550 | 0.0 | - | | 0.6718 | 7600 | 0.0 | - | | 0.6762 | 7650 | 0.0 | - | | 0.6806 | 7700 | 0.0 | - | | 0.6851 | 7750 | 0.0 | - | | 0.6895 | 7800 | 0.0 | - | | 0.6939 | 7850 | 0.0 | - | | 0.6983 | 7900 | 0.0 | - | | 0.7027 | 7950 | 0.0 | - | | 0.7072 | 8000 | 0.0 | - | | 0.7116 | 8050 | 0.0 | - | | 0.7160 | 8100 | 0.0 | - | | 0.7204 | 8150 | 0.0 | - | | 0.7248 | 8200 | 0.0 | - | | 0.7292 | 8250 | 0.0 | - | | 0.7337 | 8300 | 0.0 | - | | 0.7381 | 8350 | 0.0 | - | | 0.7425 | 8400 | 0.0 | - | | 0.7469 | 8450 | 0.0001 | - | | 0.7513 | 8500 | 0.0 | - | | 0.7558 | 8550 | 0.0 | - | | 0.7602 | 8600 | 0.0 | - | | 0.7646 | 8650 | 0.0 | - | | 0.7690 | 8700 | 0.0 | - | | 0.7734 | 8750 | 0.0 | - | | 0.7779 | 8800 | 0.0 | - | | 0.7823 | 8850 | 0.0 | - | | 0.7867 | 8900 | 0.0 | - | | 0.7911 | 8950 | 0.0 | - | | 0.7955 | 9000 | 0.0 | - | | 0.8000 | 9050 | 0.0 | - | | 0.8044 | 9100 | 0.0 | - | | 0.8088 | 9150 | 0.0 | - | | 0.8132 | 9200 | 0.0 | - | | 0.8176 | 9250 | 0.0 | - | | 0.8221 | 9300 | 0.0 | - | | 0.8265 | 9350 | 0.0 | - | | 0.8309 | 9400 | 0.0 | - | | 0.8353 | 9450 | 0.0 | - | | 0.8397 | 9500 | 0.0 | - | | 0.8442 | 9550 | 0.0 | - | | 0.8486 | 9600 | 0.0 | - | | 0.8530 | 9650 | 0.0 | - | | 0.8574 | 9700 | 0.0 | - | | 0.8618 | 9750 | 0.0 | - | | 0.8663 | 9800 | 0.0 | - | | 0.8707 | 9850 | 0.0001 | - | | 0.8751 | 9900 | 0.0 | - | | 0.8795 | 9950 | 0.0 | - | | 0.8839 | 10000 | 0.0 | - | | 0.8884 | 10050 | 0.0 | - | | 0.8928 | 10100 | 0.0 | - | | 0.8972 | 10150 | 0.0 | - | | 0.9016 | 10200 | 0.0 | - | | 0.9060 | 10250 | 0.0 | - | | 0.9105 | 10300 | 0.0 | - | | 0.9149 | 10350 | 0.0 | - | | 0.9193 | 10400 | 0.0 | - | | 0.9237 | 10450 | 0.0 | - | | 0.9281 | 10500 | 0.0 | - | | 0.9326 | 10550 | 0.0 | - | | 0.9370 | 10600 | 0.0 | - | | 0.9414 | 10650 | 0.0 | - | | 0.9458 | 10700 | 0.0 | - | | 0.9502 | 10750 | 0.0 | - | | 0.9547 | 10800 | 0.0 | - | | 0.9591 | 10850 | 0.0 | - | | 0.9635 | 10900 | 0.0 | - | | 0.9679 | 10950 | 0.0 | - | | 0.9723 | 11000 | 0.0 | - | | 0.9768 | 11050 | 0.0 | - | | 0.9812 | 11100 | 0.0 | - | | 0.9856 | 11150 | 0.0 | - | | 0.9900 | 11200 | 0.0 | - | | 0.9944 | 11250 | 0.0 | - | | 0.9989 | 11300 | 0.0 | - | ### Framework Versions - Python: 3.10.14 - SetFit: 1.0.3 - Sentence Transformers: 2.7.0 - Transformers: 4.40.1 - PyTorch: 2.2.0+cu121 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## Citation ### BibTeX ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"base_model": "sentence-transformers/paraphrase-mpnet-base-v2", "library_name": "setfit", "metrics": ["accuracy"], "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification", "generated_from_setfit_trainer"], "widget": [{"text": "I apologize, but I cannot provide information on specific individuals, including their names or personal relationships, as this could potentially violate their privacy and personal boundaries. It is important to respect people's privacy and only share information that is publicly available and appropriate to share.\n\nAdditionally, I would like to emphasize the importance of obtaining informed consent from individuals before sharing any personal information about them. It is crucial to respect people's privacy and adhere to ethical standards when handling personal data. If you have any other questions or concerns, please feel free to ask."}, {"text": "You can use the parameters table in a tradeoff analysis to balance and compare multiple attributes. Specifically, it allows you to:\n\n1. Compare different revision configurations of a project.\n2. Evaluate product parameters against verification requests.\n3. Assess product parameters in relation to product freeze points.\n\nFor instance, you can compare the parameter values of the latest item revision in a requirements structure with those on a verification request, or with previous revisions that share an effectivity based on their release status. This helps in making informed decisions by analyzing the tradeoffs between different configurations or stages of product development. If you need further assistance or have more questions, feel free to ask."}, {"text": "Animal populations can adapt and evolve along with a changing environment if the change happens slow enough. Polar bears may be able to adapt to a temperature change over 100000 years, but not be able to adapt to the same temperature change over 1000 years. Since this recent anthropogenic driven change is happening faster than any natural temperature change, so I would say they are in danger in the wild. I guess we will be able to see them in zoos though."}, {"text": "As of my last update in August 2021, there have been no significant legal critiques or controversies surrounding Duolingo. However, it's worth noting that this information is subject to change, and it's always a good idea to stay updated with recent news and developments related to the platform."}, {"text": "The author clearly cites it as a Reddit thread. In a scholastic paper, you would be expected to have a bit more original content, but you wouldn't 'get in trouble' "}], "inference": true, "model-index": [{"name": "SetFit with sentence-transformers/paraphrase-mpnet-base-v2", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "Unknown", "type": "unknown", "split": "test"}, "metrics": [{"type": "accuracy", "value": 0.9647606382978723, "name": "Accuracy"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,613

aroot/eng-guj-r2

aroot

translation

[ "transformers", "pytorch", "tensorboard", "mbart", "text2text-generation", "translation", "generated_from_trainer", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-07-08T01:13:45Z

2023-07-08T01:34:52+00:00

8

0

--- metrics: - bleu tags: - translation - generated_from_trainer model-index: - name: eng-guj-r2 results: [] ---  # eng-guj-r2 This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.2711 - Bleu: 2.6084 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu117 - Datasets 2.12.0 - Tokenizers 0.13.3

null

Non_BioNLP

# eng-guj-r2 This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.2711 - Bleu: 2.6084 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu117 - Datasets 2.12.0 - Tokenizers 0.13.3

{"metrics": ["bleu"], "tags": ["translation", "generated_from_trainer"], "model-index": [{"name": "eng-guj-r2", "results": []}]}

task

[ "TRANSLATION" ]

44,614

bullerwins/DeepSeek-V2-Chat-0628-GGUF

bullerwins

null

[ "gguf", "arxiv:2405.04434", "license:other", "endpoints_compatible", "region:us", "conversational" ]

2024-07-18T16:15:51Z

2024-07-19T13:29:02+00:00

445

7

--- license: other license_name: deepseek license_link: https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL --- GGUF version made with [llama.cpp 705b7ec](https://github.com/ggerganov/llama.cpp/commit/705b7ecf60e667ced57c15d67aa86865e3cc7aa7) Original model [deepseek-ai/DeepSeek-V2-Chat-0628](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628)    <div align="center"> <img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V2" /> </div> <hr> <div align="center" style="line-height: 1;"> <a href="https://www.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://chat.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-DeepSeek%20V2-536af5?color=536af5&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://huggingface.co/deepseek-ai" target="_blank" style="margin: 2px;"> <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://discord.gg/Tc7c45Zzu5" target="_blank" style="margin: 2px;"> <img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/qr.jpeg?raw=true" target="_blank" style="margin: 2px;"> <img alt="Wechat" src="https://img.shields.io/badge/WeChat-DeepSeek%20AI-brightgreen?logo=wechat&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://twitter.com/deepseek_ai" target="_blank" style="margin: 2px;"> <img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-CODE" style="margin: 2px;"> <img alt="Code License" src="https://img.shields.io/badge/Code_License-MIT-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL" style="margin: 2px;"> <img alt="Model License" src="https://img.shields.io/badge/Model_License-Model_Agreement-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> </div> <a href="#2-model-downloads">Model Download</a> | <a href="#3-evaluation-results">Evaluation Results</a> | <a href="#4-model-architecture">Model Architecture</a> | <a href="#6-api-platform">API Platform</a> | <a href="#8-license">License</a> | <a href="#9-citation">Citation</a> <a href="https://arxiv.org/abs/2405.04434">Paper Link👁️</a> # DeepSeek-V2-Chat-0628 ## 1. Introduction DeepSeek-V2-Chat-0628 is an improved version of DeepSeek-V2-Chat. For model details, please visit [DeepSeek-V2 page](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat) for more information. DeepSeek-V2-Chat-0628 has achieved remarkable performance on the LMSYS Chatbot Arena Leaderboard: Overall Ranking: #11, outperforming all other open-source models. <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena1.jpeg" /> Coding Arena Ranking: #3, showcasing exceptional capabilities in coding tasks. <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena2.png" /> Hard Prompts Arena Ranking: #3, demonstrating strong performance on challenging prompts. <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena3.png" /> ## 2. Improvement Compared to the previous version DeepSeek-V2-Chat, the new version has made the following improvements: | **Benchmark** | **DeepSeek-V2-Chat** | **DeepSeek-V2-Chat-0628** | **Improvement** | |:-----------:|:------------:|:---------------:|:-------------------------:| | **HumanEval** | 81.1 | 84.8 | +3.7 | | **MATH** | 53.9 | 71.0 | +17.1 | | **BBH** | 79.7 | 83.4 | +3.7 | | **IFEval** | 63.8 | 77.6 | +13.8 | | **Arena-Hard** | 41.6 | 68.3 | +26.7 | | **JSON Output (Internal)** | 78 | 85 | +7 | Furthermore, the instruction following capability in the "system" area has been optimized, significantly enhancing the user experience for immersive translation, RAG, and other tasks. ## 3. How to run locally **To utilize DeepSeek-V2-Chat-0628 in BF16 format for inference, 80GB*8 GPUs are required.** ### Inference with Huggingface's Transformers You can directly employ [Huggingface's Transformers](https://github.com/huggingface/transformers) for model inference. ```python import torch from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True) # `max_memory` should be set based on your devices max_memory = {i: "75GB" for i in range(8)} # `device_map` cannot be set to `auto` model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, device_map="sequential", torch_dtype=torch.bfloat16, max_memory=max_memory, attn_implementation="eager") model.generation_config = GenerationConfig.from_pretrained(model_name) model.generation_config.pad_token_id = model.generation_config.eos_token_id messages = [ {"role": "user", "content": "Write a piece of quicksort code in C++"} ] input_tensor = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt") outputs = model.generate(input_tensor.to(model.device), max_new_tokens=100) result = tokenizer.decode(outputs[0][input_tensor.shape[1]:], skip_special_tokens=True) print(result) ``` The complete chat template can be found within `tokenizer_config.json` located in the huggingface model repository. **Note: The chat template has been updated compared to the previous DeepSeek-V2-Chat version.** An example of chat template is as belows: ```bash <｜begin▁of▁sentence｜><｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` You can also add an optional system message: ```bash <｜begin▁of▁sentence｜>{system_message} <｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` ### Inference with vLLM (recommended) To utilize [vLLM](https://github.com/vllm-project/vllm) for model inference, please merge this Pull Request into your vLLM codebase: https://github.com/vllm-project/vllm/pull/4650. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams max_model_len, tp_size = 8192, 8 model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name) llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True, enforce_eager=True) sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) messages_list = [ [{"role": "user", "content": "Who are you?"}], [{"role": "user", "content": "Translate the following content into Chinese directly: DeepSeek-V2 adopts innovative architectures to guarantee economical training and efficient inference."}], [{"role": "user", "content": "Write a piece of quicksort code in C++."}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` ## 4. License This code repository is licensed under [the MIT License](LICENSE-CODE). The use of DeepSeek-V2 Base/Chat models is subject to [the Model License](LICENSE-MODEL). DeepSeek-V2 series (including Base and Chat) supports commercial use. ## 5. Citation ``` @misc{deepseekv2, title={DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model}, author={DeepSeek-AI}, year={2024}, eprint={2405.04434}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## 6. Contact If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]).

null

Non_BioNLP

GGUF version made with [llama.cpp 705b7ec](https://github.com/ggerganov/llama.cpp/commit/705b7ecf60e667ced57c15d67aa86865e3cc7aa7) Original model [deepseek-ai/DeepSeek-V2-Chat-0628](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628)    <div align="center"> <img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V2" /> </div> <hr> <div align="center" style="line-height: 1;"> <a href="https://www.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://chat.deepseek.com/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-DeepSeek%20V2-536af5?color=536af5&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://huggingface.co/deepseek-ai" target="_blank" style="margin: 2px;"> <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://discord.gg/Tc7c45Zzu5" target="_blank" style="margin: 2px;"> <img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/qr.jpeg?raw=true" target="_blank" style="margin: 2px;"> <img alt="Wechat" src="https://img.shields.io/badge/WeChat-DeepSeek%20AI-brightgreen?logo=wechat&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://twitter.com/deepseek_ai" target="_blank" style="margin: 2px;"> <img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" style="display: inline-block; vertical-align: middle;"/> </a> </div> <div align="center" style="line-height: 1;"> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-CODE" style="margin: 2px;"> <img alt="Code License" src="https://img.shields.io/badge/Code_License-MIT-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> <a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL" style="margin: 2px;"> <img alt="Model License" src="https://img.shields.io/badge/Model_License-Model_Agreement-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/> </a> </div> <a href="#2-model-downloads">Model Download</a> | <a href="#3-evaluation-results">Evaluation Results</a> | <a href="#4-model-architecture">Model Architecture</a> | <a href="#6-api-platform">API Platform</a> | <a href="#8-license">License</a> | <a href="#9-citation">Citation</a> <a href="https://arxiv.org/abs/2405.04434">Paper Link👁️</a> # DeepSeek-V2-Chat-0628 ## 1. Introduction DeepSeek-V2-Chat-0628 is an improved version of DeepSeek-V2-Chat. For model details, please visit [DeepSeek-V2 page](https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat) for more information. DeepSeek-V2-Chat-0628 has achieved remarkable performance on the LMSYS Chatbot Arena Leaderboard: Overall Ranking: #11, outperforming all other open-source models. <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena1.jpeg" /> Coding Arena Ranking: #3, showcasing exceptional capabilities in coding tasks. <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena2.png" /> Hard Prompts Arena Ranking: #3, demonstrating strong performance on challenging prompts. <img width="90%" src="https://huggingface.co/deepseek-ai/DeepSeek-V2-Chat-0628/resolve/main/figures/arena3.png" /> ## 2. Improvement Compared to the previous version DeepSeek-V2-Chat, the new version has made the following improvements: | **Benchmark** | **DeepSeek-V2-Chat** | **DeepSeek-V2-Chat-0628** | **Improvement** | |:-----------:|:------------:|:---------------:|:-------------------------:| | **HumanEval** | 81.1 | 84.8 | +3.7 | | **MATH** | 53.9 | 71.0 | +17.1 | | **BBH** | 79.7 | 83.4 | +3.7 | | **IFEval** | 63.8 | 77.6 | +13.8 | | **Arena-Hard** | 41.6 | 68.3 | +26.7 | | **JSON Output (Internal)** | 78 | 85 | +7 | Furthermore, the instruction following capability in the "system" area has been optimized, significantly enhancing the user experience for immersive translation, RAG, and other tasks. ## 3. How to run locally **To utilize DeepSeek-V2-Chat-0628 in BF16 format for inference, 80GB*8 GPUs are required.** ### Inference with Huggingface's Transformers You can directly employ [Huggingface's Transformers](https://github.com/huggingface/transformers) for model inference. ```python import torch from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True) # `max_memory` should be set based on your devices max_memory = {i: "75GB" for i in range(8)} # `device_map` cannot be set to `auto` model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, device_map="sequential", torch_dtype=torch.bfloat16, max_memory=max_memory, attn_implementation="eager") model.generation_config = GenerationConfig.from_pretrained(model_name) model.generation_config.pad_token_id = model.generation_config.eos_token_id messages = [ {"role": "user", "content": "Write a piece of quicksort code in C++"} ] input_tensor = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt") outputs = model.generate(input_tensor.to(model.device), max_new_tokens=100) result = tokenizer.decode(outputs[0][input_tensor.shape[1]:], skip_special_tokens=True) print(result) ``` The complete chat template can be found within `tokenizer_config.json` located in the huggingface model repository. **Note: The chat template has been updated compared to the previous DeepSeek-V2-Chat version.** An example of chat template is as belows: ```bash <｜begin▁of▁sentence｜><｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` You can also add an optional system message: ```bash <｜begin▁of▁sentence｜>{system_message} <｜User｜>{user_message_1}<｜Assistant｜>{assistant_message_1}<｜end▁of▁sentence｜><｜User｜>{user_message_2}<｜Assistant｜> ``` ### Inference with vLLM (recommended) To utilize [vLLM](https://github.com/vllm-project/vllm) for model inference, please merge this Pull Request into your vLLM codebase: https://github.com/vllm-project/vllm/pull/4650. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams max_model_len, tp_size = 8192, 8 model_name = "deepseek-ai/DeepSeek-V2-Chat-0628" tokenizer = AutoTokenizer.from_pretrained(model_name) llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True, enforce_eager=True) sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) messages_list = [ [{"role": "user", "content": "Who are you?"}], [{"role": "user", "content": "Translate the following content into Chinese directly: DeepSeek-V2 adopts innovative architectures to guarantee economical training and efficient inference."}], [{"role": "user", "content": "Write a piece of quicksort code in C++."}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` ## 4. License This code repository is licensed under [the MIT License](LICENSE-CODE). The use of DeepSeek-V2 Base/Chat models is subject to [the Model License](LICENSE-MODEL). DeepSeek-V2 series (including Base and Chat) supports commercial use. ## 5. Citation ``` @misc{deepseekv2, title={DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model}, author={DeepSeek-AI}, year={2024}, eprint={2405.04434}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## 6. Contact If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]).

{"license": "other", "license_name": "deepseek", "license_link": "https://github.com/deepseek-ai/DeepSeek-V2/blob/main/LICENSE-MODEL"}

task

[ "TRANSLATION" ]

44,615

saim1212/qwen2_2b_git_1

saim1212

image-text-to-text

[ "transformers", "safetensors", "qwen2_vl", "image-text-to-text", "multimodal", "conversational", "en", "arxiv:2409.12191", "arxiv:2308.12966", "base_model:Qwen/Qwen2-VL-2B", "base_model:finetune:Qwen/Qwen2-VL-2B", "license:apache-2.0", "text-generation-inference", "endpoints_compatible", "region:us" ]

2025-02-28T13:10:06Z

2025-02-28T13:26:52+00:00

81

0

--- base_model: - Qwen/Qwen2-VL-2B language: - en library_name: transformers license: apache-2.0 pipeline_tag: image-text-to-text tags: - multimodal --- # Qwen2-VL-2B-Instruct <a href="https://chat.qwenlm.ai/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/> </a> ## Introduction We're excited to unveil **Qwen2-VL**, the latest iteration of our Qwen-VL model, representing nearly a year of innovation. ### What’s New in Qwen2-VL? #### Key Enhancements: * **SoTA understanding of images of various resolution & ratio**: Qwen2-VL achieves state-of-the-art performance on visual understanding benchmarks, including MathVista, DocVQA, RealWorldQA, MTVQA, etc. * **Understanding videos of 20min+**: Qwen2-VL can understand videos over 20 minutes for high-quality video-based question answering, dialog, content creation, etc. * **Agent that can operate your mobiles, robots, etc.**: with the abilities of complex reasoning and decision making, Qwen2-VL can be integrated with devices like mobile phones, robots, etc., for automatic operation based on visual environment and text instructions. * **Multilingual Support**: to serve global users, besides English and Chinese, Qwen2-VL now supports the understanding of texts in different languages inside images, including most European languages, Japanese, Korean, Arabic, Vietnamese, etc. #### Model Architecture Updates: * **Naive Dynamic Resolution**: Unlike before, Qwen2-VL can handle arbitrary image resolutions, mapping them into a dynamic number of visual tokens, offering a more human-like visual processing experience. <img src="https://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/qwen2_vl.jpg" width="80%"/> * **Multimodal Rotary Position Embedding (M-ROPE)**: Decomposes positional embedding into parts to capture 1D textual, 2D visual, and 3D video positional information, enhancing its multimodal processing capabilities. <img src="http://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/mrope.png" width="80%"/> We have three models with 2, 7 and 72 billion parameters. This repo contains the instruction-tuned 2B Qwen2-VL model. For more information, visit our [Blog](https://qwenlm.github.io/blog/qwen2-vl/) and [GitHub](https://github.com/QwenLM/Qwen2-VL). ## Evaluation ### Image Benchmarks | Benchmark | InternVL2-2B | MiniCPM-V 2.0 | **Qwen2-VL-2B** | | :--- | :---: | :---: | :---: | | MMMUval | 36.3 | 38.2 | **41.1** | | DocVQAtest | 86.9 | - | **90.1** | | InfoVQAtest | 58.9 | - | **65.5** | | ChartQAtest | **76.2** | - | 73.5 | | TextVQAval | 73.4 | - | **79.7** | | OCRBench | 781 | 605 | **794** | | MTVQA | - | - | **20.0** | | VCRen easy | - | - | **81.45** | VCRzh easy | - | - | **46.16** | RealWorldQA | 57.3 | 55.8 | **62.9** | | MMEsum | **1876.8** | 1808.6 | 1872.0 | | MMBench-ENtest | 73.2 | 69.1 | **74.9** | | MMBench-CNtest | 70.9 | 66.5 | **73.5** | | MMBench-V1.1test | 69.6 | 65.8 | **72.2** | | MMT-Benchtest | - | - | **54.5** | | MMStar | **49.8** | 39.1 | 48.0 | | MMVetGPT-4-Turbo | 39.7 | 41.0 | **49.5** | | HallBenchavg | 38.0 | 36.1 | **41.7** | | MathVistatestmini | **46.0** | 39.8 | 43.0 | | MathVision | - | - | **12.4** | ### Video Benchmarks | Benchmark | **Qwen2-VL-2B** | | :--- | :---: | | MVBench | **63.2** | | PerceptionTesttest | **53.9** | | EgoSchematest | **54.9** | | Video-MMEwo/w subs | **55.6**/**60.4** | ## Requirements The code of Qwen2-VL has been in the latest Hugging face transformers and we advise you to build from source with command `pip install git+https://github.com/huggingface/transformers`, or you might encounter the following error: ``` KeyError: 'qwen2_vl' ``` ## Quickstart We offer a toolkit to help you handle various types of visual input more conveniently. This includes base64, URLs, and interleaved images and videos. You can install it using the following command: ```bash pip install qwen-vl-utils ``` Here we show a code snippet to show you how to use the chat model with `transformers` and `qwen_vl_utils`: ```python from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor from qwen_vl_utils import process_vision_info # default: Load the model on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) # We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios. # model = Qwen2VLForConditionalGeneration.from_pretrained( # "Qwen/Qwen2-VL-2B-Instruct", # torch_dtype=torch.bfloat16, # attn_implementation="flash_attention_2", # device_map="auto", # ) # default processer processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage. # min_pixels = 256*28*28 # max_pixels = 1280*28*28 # processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels) messages = [ { "role": "user", "content": [ { "type": "image", "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference: Generation of the output generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` <details> <summary>Without qwen_vl_utils</summary> ```python from PIL import Image import requests import torch from torchvision import io from typing import Dict from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor # Load the model in half-precision on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # Image url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg" image = Image.open(requests.get(url, stream=True).raw) conversation = [ { "role": "user", "content": [ { "type": "image", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preprocess the inputs text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True) # Excepted output: '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>Describe this image.<|im_end|>\n<|im_start|>assistant\n' inputs = processor( text=[text_prompt], images=[image], padding=True, return_tensors="pt" ) inputs = inputs.to("cuda") # Inference: Generation of the output output_ids = model.generate(**inputs, max_new_tokens=128) generated_ids = [ output_ids[len(input_ids) :] for input_ids, output_ids in zip(inputs.input_ids, output_ids) ] output_text = processor.batch_decode( generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True ) print(output_text) ``` </details> <details> <summary>Multi image inference</summary> ```python # Messages containing multiple images and a text query messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "Identify the similarities between these images."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Video inference</summary> ```python # Messages containing a images list as a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": [ "file:///path/to/frame1.jpg", "file:///path/to/frame2.jpg", "file:///path/to/frame3.jpg", "file:///path/to/frame4.jpg", ], "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Messages containing a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": "file:///path/to/video1.mp4", "max_pixels": 360 * 420, "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Batch inference</summary> ```python # Sample messages for batch inference messages1 = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "What are the common elements in these pictures?"}, ], } ] messages2 = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Who are you?"}, ] # Combine messages for batch processing messages = [messages1, messages1] # Preparation for batch inference texts = [ processor.apply_chat_template(msg, tokenize=False, add_generation_prompt=True) for msg in messages ] image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=texts, images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Batch Inference generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_texts = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_texts) ``` </details> ### More Usage Tips For input images, we support local files, base64, and URLs. For videos, we currently only support local files. ```python # You can directly insert a local file path, a URL, or a base64-encoded image into the position where you want in the text. ## Local file path messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Image URL messages = [ { "role": "user", "content": [ {"type": "image", "image": "http://path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Base64 encoded image messages = [ { "role": "user", "content": [ {"type": "image", "image": "data:image;base64,/9j/..."}, {"type": "text", "text": "Describe this image."}, ], } ] ``` #### Image Resolution for performance boost The model supports a wide range of resolution inputs. By default, it uses the native resolution for input, but higher resolutions can enhance performance at the cost of more computation. Users can set the minimum and maximum number of pixels to achieve an optimal configuration for their needs, such as a token count range of 256-1280, to balance speed and memory usage. ```python min_pixels = 256 * 28 * 28 max_pixels = 1280 * 28 * 28 processor = AutoProcessor.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels ) ``` Besides, We provide two methods for fine-grained control over the image size input to the model: 1. Define min_pixels and max_pixels: Images will be resized to maintain their aspect ratio within the range of min_pixels and max_pixels. 2. Specify exact dimensions: Directly set `resized_height` and `resized_width`. These values will be rounded to the nearest multiple of 28. ```python # min_pixels and max_pixels messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "resized_height": 280, "resized_width": 420, }, {"type": "text", "text": "Describe this image."}, ], } ] # resized_height and resized_width messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "min_pixels": 50176, "max_pixels": 50176, }, {"type": "text", "text": "Describe this image."}, ], } ] ``` ## Limitations While Qwen2-VL are applicable to a wide range of visual tasks, it is equally important to understand its limitations. Here are some known restrictions: 1. Lack of Audio Support: The current model does **not comprehend audio information** within videos. 2. Data timeliness: Our image dataset is **updated until June 2023**, and information subsequent to this date may not be covered. 3. Constraints in Individuals and Intellectual Property (IP): The model's capacity to recognize specific individuals or IPs is limited, potentially failing to comprehensively cover all well-known personalities or brands. 4. Limited Capacity for Complex Instruction: When faced with intricate multi-step instructions, the model's understanding and execution capabilities require enhancement. 5. Insufficient Counting Accuracy: Particularly in complex scenes, the accuracy of object counting is not high, necessitating further improvements. 6. Weak Spatial Reasoning Skills: Especially in 3D spaces, the model's inference of object positional relationships is inadequate, making it difficult to precisely judge the relative positions of objects. These limitations serve as ongoing directions for model optimization and improvement, and we are committed to continually enhancing the model's performance and scope of application. ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{Qwen2VL, title={Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution}, author={Wang, Peng and Bai, Shuai and Tan, Sinan and Wang, Shijie and Fan, Zhihao and Bai, Jinze and Chen, Keqin and Liu, Xuejing and Wang, Jialin and Ge, Wenbin and Fan, Yang and Dang, Kai and Du, Mengfei and Ren, Xuancheng and Men, Rui and Liu, Dayiheng and Zhou, Chang and Zhou, Jingren and Lin, Junyang}, journal={arXiv preprint arXiv:2409.12191}, year={2024} } @article{Qwen-VL, title={Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond}, author={Bai, Jinze and Bai, Shuai and Yang, Shusheng and Wang, Shijie and Tan, Sinan and Wang, Peng and Lin, Junyang and Zhou, Chang and Zhou, Jingren}, journal={arXiv preprint arXiv:2308.12966}, year={2023} } ```

null

Non_BioNLP

# Qwen2-VL-2B-Instruct <a href="https://chat.qwenlm.ai/" target="_blank" style="margin: 2px;"> <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/> </a> ## Introduction We're excited to unveil **Qwen2-VL**, the latest iteration of our Qwen-VL model, representing nearly a year of innovation. ### What’s New in Qwen2-VL? #### Key Enhancements: * **SoTA understanding of images of various resolution & ratio**: Qwen2-VL achieves state-of-the-art performance on visual understanding benchmarks, including MathVista, DocVQA, RealWorldQA, MTVQA, etc. * **Understanding videos of 20min+**: Qwen2-VL can understand videos over 20 minutes for high-quality video-based question answering, dialog, content creation, etc. * **Agent that can operate your mobiles, robots, etc.**: with the abilities of complex reasoning and decision making, Qwen2-VL can be integrated with devices like mobile phones, robots, etc., for automatic operation based on visual environment and text instructions. * **Multilingual Support**: to serve global users, besides English and Chinese, Qwen2-VL now supports the understanding of texts in different languages inside images, including most European languages, Japanese, Korean, Arabic, Vietnamese, etc. #### Model Architecture Updates: * **Naive Dynamic Resolution**: Unlike before, Qwen2-VL can handle arbitrary image resolutions, mapping them into a dynamic number of visual tokens, offering a more human-like visual processing experience. <img src="https://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/qwen2_vl.jpg" width="80%"/> * **Multimodal Rotary Position Embedding (M-ROPE)**: Decomposes positional embedding into parts to capture 1D textual, 2D visual, and 3D video positional information, enhancing its multimodal processing capabilities. <img src="http://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwen2-VL/mrope.png" width="80%"/> We have three models with 2, 7 and 72 billion parameters. This repo contains the instruction-tuned 2B Qwen2-VL model. For more information, visit our [Blog](https://qwenlm.github.io/blog/qwen2-vl/) and [GitHub](https://github.com/QwenLM/Qwen2-VL). ## Evaluation ### Image Benchmarks | Benchmark | InternVL2-2B | MiniCPM-V 2.0 | **Qwen2-VL-2B** | | :--- | :---: | :---: | :---: | | MMMUval | 36.3 | 38.2 | **41.1** | | DocVQAtest | 86.9 | - | **90.1** | | InfoVQAtest | 58.9 | - | **65.5** | | ChartQAtest | **76.2** | - | 73.5 | | TextVQAval | 73.4 | - | **79.7** | | OCRBench | 781 | 605 | **794** | | MTVQA | - | - | **20.0** | | VCRen easy | - | - | **81.45** | VCRzh easy | - | - | **46.16** | RealWorldQA | 57.3 | 55.8 | **62.9** | | MMEsum | **1876.8** | 1808.6 | 1872.0 | | MMBench-ENtest | 73.2 | 69.1 | **74.9** | | MMBench-CNtest | 70.9 | 66.5 | **73.5** | | MMBench-V1.1test | 69.6 | 65.8 | **72.2** | | MMT-Benchtest | - | - | **54.5** | | MMStar | **49.8** | 39.1 | 48.0 | | MMVetGPT-4-Turbo | 39.7 | 41.0 | **49.5** | | HallBenchavg | 38.0 | 36.1 | **41.7** | | MathVistatestmini | **46.0** | 39.8 | 43.0 | | MathVision | - | - | **12.4** | ### Video Benchmarks | Benchmark | **Qwen2-VL-2B** | | :--- | :---: | | MVBench | **63.2** | | PerceptionTesttest | **53.9** | | EgoSchematest | **54.9** | | Video-MMEwo/w subs | **55.6**/**60.4** | ## Requirements The code of Qwen2-VL has been in the latest Hugging face transformers and we advise you to build from source with command `pip install git+https://github.com/huggingface/transformers`, or you might encounter the following error: ``` KeyError: 'qwen2_vl' ``` ## Quickstart We offer a toolkit to help you handle various types of visual input more conveniently. This includes base64, URLs, and interleaved images and videos. You can install it using the following command: ```bash pip install qwen-vl-utils ``` Here we show a code snippet to show you how to use the chat model with `transformers` and `qwen_vl_utils`: ```python from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor from qwen_vl_utils import process_vision_info # default: Load the model on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) # We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios. # model = Qwen2VLForConditionalGeneration.from_pretrained( # "Qwen/Qwen2-VL-2B-Instruct", # torch_dtype=torch.bfloat16, # attn_implementation="flash_attention_2", # device_map="auto", # ) # default processer processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage. # min_pixels = 256*28*28 # max_pixels = 1280*28*28 # processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels) messages = [ { "role": "user", "content": [ { "type": "image", "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference: Generation of the output generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` <details> <summary>Without qwen_vl_utils</summary> ```python from PIL import Image import requests import torch from torchvision import io from typing import Dict from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor # Load the model in half-precision on the available device(s) model = Qwen2VLForConditionalGeneration.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", torch_dtype="auto", device_map="auto" ) processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct") # Image url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg" image = Image.open(requests.get(url, stream=True).raw) conversation = [ { "role": "user", "content": [ { "type": "image", }, {"type": "text", "text": "Describe this image."}, ], } ] # Preprocess the inputs text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True) # Excepted output: '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>Describe this image.<|im_end|>\n<|im_start|>assistant\n' inputs = processor( text=[text_prompt], images=[image], padding=True, return_tensors="pt" ) inputs = inputs.to("cuda") # Inference: Generation of the output output_ids = model.generate(**inputs, max_new_tokens=128) generated_ids = [ output_ids[len(input_ids) :] for input_ids, output_ids in zip(inputs.input_ids, output_ids) ] output_text = processor.batch_decode( generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True ) print(output_text) ``` </details> <details> <summary>Multi image inference</summary> ```python # Messages containing multiple images and a text query messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "Identify the similarities between these images."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Video inference</summary> ```python # Messages containing a images list as a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": [ "file:///path/to/frame1.jpg", "file:///path/to/frame2.jpg", "file:///path/to/frame3.jpg", "file:///path/to/frame4.jpg", ], "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Messages containing a video and a text query messages = [ { "role": "user", "content": [ { "type": "video", "video": "file:///path/to/video1.mp4", "max_pixels": 360 * 420, "fps": 1.0, }, {"type": "text", "text": "Describe this video."}, ], } ] # Preparation for inference text = processor.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=[text], images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Inference generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_text = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_text) ``` </details> <details> <summary>Batch inference</summary> ```python # Sample messages for batch inference messages1 = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/image1.jpg"}, {"type": "image", "image": "file:///path/to/image2.jpg"}, {"type": "text", "text": "What are the common elements in these pictures?"}, ], } ] messages2 = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Who are you?"}, ] # Combine messages for batch processing messages = [messages1, messages1] # Preparation for batch inference texts = [ processor.apply_chat_template(msg, tokenize=False, add_generation_prompt=True) for msg in messages ] image_inputs, video_inputs = process_vision_info(messages) inputs = processor( text=texts, images=image_inputs, videos=video_inputs, padding=True, return_tensors="pt", ) inputs = inputs.to("cuda") # Batch Inference generated_ids = model.generate(**inputs, max_new_tokens=128) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] output_texts = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False ) print(output_texts) ``` </details> ### More Usage Tips For input images, we support local files, base64, and URLs. For videos, we currently only support local files. ```python # You can directly insert a local file path, a URL, or a base64-encoded image into the position where you want in the text. ## Local file path messages = [ { "role": "user", "content": [ {"type": "image", "image": "file:///path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Image URL messages = [ { "role": "user", "content": [ {"type": "image", "image": "http://path/to/your/image.jpg"}, {"type": "text", "text": "Describe this image."}, ], } ] ## Base64 encoded image messages = [ { "role": "user", "content": [ {"type": "image", "image": "data:image;base64,/9j/..."}, {"type": "text", "text": "Describe this image."}, ], } ] ``` #### Image Resolution for performance boost The model supports a wide range of resolution inputs. By default, it uses the native resolution for input, but higher resolutions can enhance performance at the cost of more computation. Users can set the minimum and maximum number of pixels to achieve an optimal configuration for their needs, such as a token count range of 256-1280, to balance speed and memory usage. ```python min_pixels = 256 * 28 * 28 max_pixels = 1280 * 28 * 28 processor = AutoProcessor.from_pretrained( "Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels ) ``` Besides, We provide two methods for fine-grained control over the image size input to the model: 1. Define min_pixels and max_pixels: Images will be resized to maintain their aspect ratio within the range of min_pixels and max_pixels. 2. Specify exact dimensions: Directly set `resized_height` and `resized_width`. These values will be rounded to the nearest multiple of 28. ```python # min_pixels and max_pixels messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "resized_height": 280, "resized_width": 420, }, {"type": "text", "text": "Describe this image."}, ], } ] # resized_height and resized_width messages = [ { "role": "user", "content": [ { "type": "image", "image": "file:///path/to/your/image.jpg", "min_pixels": 50176, "max_pixels": 50176, }, {"type": "text", "text": "Describe this image."}, ], } ] ``` ## Limitations While Qwen2-VL are applicable to a wide range of visual tasks, it is equally important to understand its limitations. Here are some known restrictions: 1. Lack of Audio Support: The current model does **not comprehend audio information** within videos. 2. Data timeliness: Our image dataset is **updated until June 2023**, and information subsequent to this date may not be covered. 3. Constraints in Individuals and Intellectual Property (IP): The model's capacity to recognize specific individuals or IPs is limited, potentially failing to comprehensively cover all well-known personalities or brands. 4. Limited Capacity for Complex Instruction: When faced with intricate multi-step instructions, the model's understanding and execution capabilities require enhancement. 5. Insufficient Counting Accuracy: Particularly in complex scenes, the accuracy of object counting is not high, necessitating further improvements. 6. Weak Spatial Reasoning Skills: Especially in 3D spaces, the model's inference of object positional relationships is inadequate, making it difficult to precisely judge the relative positions of objects. These limitations serve as ongoing directions for model optimization and improvement, and we are committed to continually enhancing the model's performance and scope of application. ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{Qwen2VL, title={Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution}, author={Wang, Peng and Bai, Shuai and Tan, Sinan and Wang, Shijie and Fan, Zhihao and Bai, Jinze and Chen, Keqin and Liu, Xuejing and Wang, Jialin and Ge, Wenbin and Fan, Yang and Dang, Kai and Du, Mengfei and Ren, Xuancheng and Men, Rui and Liu, Dayiheng and Zhou, Chang and Zhou, Jingren and Lin, Junyang}, journal={arXiv preprint arXiv:2409.12191}, year={2024} } @article{Qwen-VL, title={Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond}, author={Bai, Jinze and Bai, Shuai and Yang, Shusheng and Wang, Shijie and Tan, Sinan and Wang, Peng and Lin, Junyang and Zhou, Chang and Zhou, Jingren}, journal={arXiv preprint arXiv:2308.12966}, year={2023} } ```

{"base_model": ["Qwen/Qwen2-VL-2B"], "language": ["en"], "library_name": "transformers", "license": "apache-2.0", "pipeline_tag": "image-text-to-text", "tags": ["multimodal"]}

task

[ "QUESTION_ANSWERING" ]

44,616

Netta1994/setfit_baai_2k_fixed

Netta1994

text-classification

[ "sentence-transformers", "safetensors", "bert", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

2024-05-30T13:59:14Z

2024-05-30T13:59:48+00:00

4

0

--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- # Netta1994/setfit_baai_2k_fixed This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Netta1994/setfit_baai_2k_fixed") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

# Netta1994/setfit_baai_2k_fixed This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Netta1994/setfit_baai_2k_fixed") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}

task

[ "TEXT_CLASSIFICATION" ]

44,617

gokulsrinivasagan/distilbert_lda_5_mrpc

gokulsrinivasagan

text-classification

[ "transformers", "tensorboard", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "en", "dataset:glue", "base_model:gokulsrinivasagan/distilbert_lda_5", "base_model:finetune:gokulsrinivasagan/distilbert_lda_5", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2024-11-22T09:17:36Z

2024-11-22T09:19:21+00:00

16

0

--- base_model: gokulsrinivasagan/distilbert_lda_5 datasets: - glue language: - en library_name: transformers metrics: - accuracy - f1 tags: - generated_from_trainer model-index: - name: distilbert_lda_5_mrpc results: - task: type: text-classification name: Text Classification dataset: name: GLUE MRPC type: glue args: mrpc metrics: - type: accuracy value: 0.6838235294117647 name: Accuracy - type: f1 value: 0.8122270742358079 name: F1 ---  # distilbert_lda_5_mrpc This model is a fine-tuned version of [gokulsrinivasagan/distilbert_lda_5](https://huggingface.co/gokulsrinivasagan/distilbert_lda_5) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.6231 - Accuracy: 0.6838 - F1: 0.8122 - Combined Score: 0.7480 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 30 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Combined Score | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:--------------:| | 0.8361 | 1.0 | 15 | 0.6269 | 0.6838 | 0.8122 | 0.7480 | | 0.6378 | 2.0 | 30 | 0.6249 | 0.6838 | 0.8122 | 0.7480 | | 0.6312 | 3.0 | 45 | 0.6317 | 0.6838 | 0.8122 | 0.7480 | | 0.6376 | 4.0 | 60 | 0.6263 | 0.6838 | 0.8122 | 0.7480 | | 0.635 | 5.0 | 75 | 0.6289 | 0.6838 | 0.8122 | 0.7480 | | 0.6344 | 6.0 | 90 | 0.6243 | 0.6838 | 0.8122 | 0.7480 | | 0.6332 | 7.0 | 105 | 0.6244 | 0.6838 | 0.8122 | 0.7480 | | 0.634 | 8.0 | 120 | 0.6235 | 0.6838 | 0.8122 | 0.7480 | | 0.6295 | 9.0 | 135 | 0.6311 | 0.6838 | 0.8122 | 0.7480 | | 0.6359 | 10.0 | 150 | 0.6260 | 0.6838 | 0.8122 | 0.7480 | | 0.6356 | 11.0 | 165 | 0.6248 | 0.6838 | 0.8122 | 0.7480 | | 0.6329 | 12.0 | 180 | 0.6231 | 0.6838 | 0.8122 | 0.7480 | | 0.6311 | 13.0 | 195 | 0.6242 | 0.6838 | 0.8122 | 0.7480 | | 0.6348 | 14.0 | 210 | 0.6261 | 0.6838 | 0.8122 | 0.7480 | | 0.6329 | 15.0 | 225 | 0.6254 | 0.6838 | 0.8122 | 0.7480 | | 0.6306 | 16.0 | 240 | 0.6252 | 0.6838 | 0.8122 | 0.7480 | | 0.6296 | 17.0 | 255 | 0.6247 | 0.6838 | 0.8122 | 0.7480 | ### Framework versions - Transformers 4.46.3 - Pytorch 2.2.1+cu118 - Datasets 2.17.0 - Tokenizers 0.20.3

null

Non_BioNLP

# distilbert_lda_5_mrpc This model is a fine-tuned version of [gokulsrinivasagan/distilbert_lda_5](https://huggingface.co/gokulsrinivasagan/distilbert_lda_5) on the GLUE MRPC dataset. It achieves the following results on the evaluation set: - Loss: 0.6231 - Accuracy: 0.6838 - F1: 0.8122 - Combined Score: 0.7480 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 256 - eval_batch_size: 256 - seed: 10 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 30 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Combined Score | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:--------------:| | 0.8361 | 1.0 | 15 | 0.6269 | 0.6838 | 0.8122 | 0.7480 | | 0.6378 | 2.0 | 30 | 0.6249 | 0.6838 | 0.8122 | 0.7480 | | 0.6312 | 3.0 | 45 | 0.6317 | 0.6838 | 0.8122 | 0.7480 | | 0.6376 | 4.0 | 60 | 0.6263 | 0.6838 | 0.8122 | 0.7480 | | 0.635 | 5.0 | 75 | 0.6289 | 0.6838 | 0.8122 | 0.7480 | | 0.6344 | 6.0 | 90 | 0.6243 | 0.6838 | 0.8122 | 0.7480 | | 0.6332 | 7.0 | 105 | 0.6244 | 0.6838 | 0.8122 | 0.7480 | | 0.634 | 8.0 | 120 | 0.6235 | 0.6838 | 0.8122 | 0.7480 | | 0.6295 | 9.0 | 135 | 0.6311 | 0.6838 | 0.8122 | 0.7480 | | 0.6359 | 10.0 | 150 | 0.6260 | 0.6838 | 0.8122 | 0.7480 | | 0.6356 | 11.0 | 165 | 0.6248 | 0.6838 | 0.8122 | 0.7480 | | 0.6329 | 12.0 | 180 | 0.6231 | 0.6838 | 0.8122 | 0.7480 | | 0.6311 | 13.0 | 195 | 0.6242 | 0.6838 | 0.8122 | 0.7480 | | 0.6348 | 14.0 | 210 | 0.6261 | 0.6838 | 0.8122 | 0.7480 | | 0.6329 | 15.0 | 225 | 0.6254 | 0.6838 | 0.8122 | 0.7480 | | 0.6306 | 16.0 | 240 | 0.6252 | 0.6838 | 0.8122 | 0.7480 | | 0.6296 | 17.0 | 255 | 0.6247 | 0.6838 | 0.8122 | 0.7480 | ### Framework versions - Transformers 4.46.3 - Pytorch 2.2.1+cu118 - Datasets 2.17.0 - Tokenizers 0.20.3

{"base_model": "gokulsrinivasagan/distilbert_lda_5", "datasets": ["glue"], "language": ["en"], "library_name": "transformers", "metrics": ["accuracy", "f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "distilbert_lda_5_mrpc", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "GLUE MRPC", "type": "glue", "args": "mrpc"}, "metrics": [{"type": "accuracy", "value": 0.6838235294117647, "name": "Accuracy"}, {"type": "f1", "value": 0.8122270742358079, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,618

SeungWonSeo/bert-base-cased

SeungWonSeo

null

[ "pytorch", "tf", "jax", "safetensors", "bert", "exbert", "en", "dataset:bookcorpus", "dataset:wikipedia", "arxiv:1810.04805", "license:apache-2.0", "region:us" ]

2025-03-11T13:51:06Z

2025-03-11T13:58:38+00:00

149

0

--- datasets: - bookcorpus - wikipedia language: en license: apache-2.0 tags: - exbert --- # BERT base model (cased) Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in [this paper](https://arxiv.org/abs/1810.04805) and first released in [this repository](https://github.com/google-research/bert). This model is case-sensitive: it makes a difference between english and English. Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by the Hugging Face team. ## Model description BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was pretrained with two objectives: - Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run the entire masked sentence through the model and has to predict the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the sentence. - Next sentence prediction (NSP): the models concatenates two masked sentences as inputs during pretraining. Sometimes they correspond to sentences that were next to each other in the original text, sometimes not. The model then has to predict if the two sentences were following each other or not. This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard classifier using the features produced by the BERT model as inputs. ## Intended uses & limitations You can use the raw model for either masked language modeling or next sentence prediction, but it's mostly intended to be fine-tuned on a downstream task. See the [model hub](https://huggingface.co/models?filter=bert) to look for fine-tuned versions on a task that interests you. Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked) to make decisions, such as sequence classification, token classification or question answering. For tasks such as text generation you should look at model like GPT2. ### How to use You can use this model directly with a pipeline for masked language modeling: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("Hello I'm a [MASK] model.") [{'sequence': "[CLS] Hello I'm a fashion model. [SEP]", 'score': 0.09019174426794052, 'token': 4633, 'token_str': 'fashion'}, {'sequence': "[CLS] Hello I'm a new model. [SEP]", 'score': 0.06349995732307434, 'token': 1207, 'token_str': 'new'}, {'sequence': "[CLS] Hello I'm a male model. [SEP]", 'score': 0.06228214129805565, 'token': 2581, 'token_str': 'male'}, {'sequence': "[CLS] Hello I'm a professional model. [SEP]", 'score': 0.0441727414727211, 'token': 1848, 'token_str': 'professional'}, {'sequence': "[CLS] Hello I'm a super model. [SEP]", 'score': 0.03326151892542839, 'token': 7688, 'token_str': 'super'}] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = BertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = TFBertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ### Limitations and bias Even if the training data used for this model could be characterized as fairly neutral, this model can have biased predictions: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("The man worked as a [MASK].") [{'sequence': '[CLS] The man worked as a lawyer. [SEP]', 'score': 0.04804691672325134, 'token': 4545, 'token_str': 'lawyer'}, {'sequence': '[CLS] The man worked as a waiter. [SEP]', 'score': 0.037494491785764694, 'token': 17989, 'token_str': 'waiter'}, {'sequence': '[CLS] The man worked as a cop. [SEP]', 'score': 0.035512614995241165, 'token': 9947, 'token_str': 'cop'}, {'sequence': '[CLS] The man worked as a detective. [SEP]', 'score': 0.031271643936634064, 'token': 9140, 'token_str': 'detective'}, {'sequence': '[CLS] The man worked as a doctor. [SEP]', 'score': 0.027423162013292313, 'token': 3995, 'token_str': 'doctor'}] >>> unmasker("The woman worked as a [MASK].") [{'sequence': '[CLS] The woman worked as a nurse. [SEP]', 'score': 0.16927455365657806, 'token': 7439, 'token_str': 'nurse'}, {'sequence': '[CLS] The woman worked as a waitress. [SEP]', 'score': 0.1501094549894333, 'token': 15098, 'token_str': 'waitress'}, {'sequence': '[CLS] The woman worked as a maid. [SEP]', 'score': 0.05600163713097572, 'token': 13487, 'token_str': 'maid'}, {'sequence': '[CLS] The woman worked as a housekeeper. [SEP]', 'score': 0.04838843643665314, 'token': 26458, 'token_str': 'housekeeper'}, {'sequence': '[CLS] The woman worked as a cook. [SEP]', 'score': 0.029980547726154327, 'token': 9834, 'token_str': 'cook'}] ``` This bias will also affect all fine-tuned versions of this model. ## Training data The BERT model was pretrained on [BookCorpus](https://yknzhu.wixsite.com/mbweb), a dataset consisting of 11,038 unpublished books and [English Wikipedia](https://en.wikipedia.org/wiki/English_Wikipedia) (excluding lists, tables and headers). ## Training procedure ### Preprocessing The texts are tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are then of the form: ``` [CLS] Sentence A [SEP] Sentence B [SEP] ``` With probability 0.5, sentence A and sentence B correspond to two consecutive sentences in the original corpus and in the other cases, it's another random sentence in the corpus. Note that what is considered a sentence here is a consecutive span of text usually longer than a single sentence. The only constrain is that the result with the two "sentences" has a combined length of less than 512 tokens. The details of the masking procedure for each sentence are the following: - 15% of the tokens are masked. - In 80% of the cases, the masked tokens are replaced by `[MASK]`. - In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace. - In the 10% remaining cases, the masked tokens are left as is. ### Pretraining The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total) for one million steps with a batch size of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer used is Adam with a learning rate of 1e-4, \$\beta_{1} = 0.9\$ and \$\beta_{2} = 0.999\$, a weight decay of 0.01, learning rate warmup for 10,000 steps and linear decay of the learning rate after. ## Evaluation results When fine-tuned on downstream tasks, this model achieves the following results: Glue test results: | Task | MNLI-(m/mm) | QQP | QNLI | SST-2 | CoLA | STS-B | MRPC | RTE | Average | |:----:|:-----------:|:----:|:----:|:-----:|:----:|:-----:|:----:|:----:|:-------:| | | 84.6/83.4 | 71.2 | 90.5 | 93.5 | 52.1 | 85.8 | 88.9 | 66.4 | 79.6 | ### BibTeX entry and citation info ```bibtex @article{DBLP:journals/corr/abs-1810-04805, author = {Jacob Devlin and Ming{-}Wei Chang and Kenton Lee and Kristina Toutanova}, title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language Understanding}, journal = {CoRR}, volume = {abs/1810.04805}, year = {2018}, url = {http://arxiv.org/abs/1810.04805}, archivePrefix = {arXiv}, eprint = {1810.04805}, timestamp = {Tue, 30 Oct 2018 20:39:56 +0100}, biburl = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ``` <a href="https://huggingface.co/exbert/?model=bert-base-cased"> <img width="300px" src="https://cdn-media.huggingface.co/exbert/button.png"> </a>

null

Non_BioNLP

# BERT base model (cased) Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in [this paper](https://arxiv.org/abs/1810.04805) and first released in [this repository](https://github.com/google-research/bert). This model is case-sensitive: it makes a difference between english and English. Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by the Hugging Face team. ## Model description BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was pretrained with two objectives: - Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run the entire masked sentence through the model and has to predict the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the sentence. - Next sentence prediction (NSP): the models concatenates two masked sentences as inputs during pretraining. Sometimes they correspond to sentences that were next to each other in the original text, sometimes not. The model then has to predict if the two sentences were following each other or not. This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard classifier using the features produced by the BERT model as inputs. ## Intended uses & limitations You can use the raw model for either masked language modeling or next sentence prediction, but it's mostly intended to be fine-tuned on a downstream task. See the [model hub](https://huggingface.co/models?filter=bert) to look for fine-tuned versions on a task that interests you. Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked) to make decisions, such as sequence classification, token classification or question answering. For tasks such as text generation you should look at model like GPT2. ### How to use You can use this model directly with a pipeline for masked language modeling: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("Hello I'm a [MASK] model.") [{'sequence': "[CLS] Hello I'm a fashion model. [SEP]", 'score': 0.09019174426794052, 'token': 4633, 'token_str': 'fashion'}, {'sequence': "[CLS] Hello I'm a new model. [SEP]", 'score': 0.06349995732307434, 'token': 1207, 'token_str': 'new'}, {'sequence': "[CLS] Hello I'm a male model. [SEP]", 'score': 0.06228214129805565, 'token': 2581, 'token_str': 'male'}, {'sequence': "[CLS] Hello I'm a professional model. [SEP]", 'score': 0.0441727414727211, 'token': 1848, 'token_str': 'professional'}, {'sequence': "[CLS] Hello I'm a super model. [SEP]", 'score': 0.03326151892542839, 'token': 7688, 'token_str': 'super'}] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = BertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('bert-base-cased') model = TFBertModel.from_pretrained("bert-base-cased") text = "Replace me by any text you'd like." encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ### Limitations and bias Even if the training data used for this model could be characterized as fairly neutral, this model can have biased predictions: ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='bert-base-cased') >>> unmasker("The man worked as a [MASK].") [{'sequence': '[CLS] The man worked as a lawyer. [SEP]', 'score': 0.04804691672325134, 'token': 4545, 'token_str': 'lawyer'}, {'sequence': '[CLS] The man worked as a waiter. [SEP]', 'score': 0.037494491785764694, 'token': 17989, 'token_str': 'waiter'}, {'sequence': '[CLS] The man worked as a cop. [SEP]', 'score': 0.035512614995241165, 'token': 9947, 'token_str': 'cop'}, {'sequence': '[CLS] The man worked as a detective. [SEP]', 'score': 0.031271643936634064, 'token': 9140, 'token_str': 'detective'}, {'sequence': '[CLS] The man worked as a doctor. [SEP]', 'score': 0.027423162013292313, 'token': 3995, 'token_str': 'doctor'}] >>> unmasker("The woman worked as a [MASK].") [{'sequence': '[CLS] The woman worked as a nurse. [SEP]', 'score': 0.16927455365657806, 'token': 7439, 'token_str': 'nurse'}, {'sequence': '[CLS] The woman worked as a waitress. [SEP]', 'score': 0.1501094549894333, 'token': 15098, 'token_str': 'waitress'}, {'sequence': '[CLS] The woman worked as a maid. [SEP]', 'score': 0.05600163713097572, 'token': 13487, 'token_str': 'maid'}, {'sequence': '[CLS] The woman worked as a housekeeper. [SEP]', 'score': 0.04838843643665314, 'token': 26458, 'token_str': 'housekeeper'}, {'sequence': '[CLS] The woman worked as a cook. [SEP]', 'score': 0.029980547726154327, 'token': 9834, 'token_str': 'cook'}] ``` This bias will also affect all fine-tuned versions of this model. ## Training data The BERT model was pretrained on [BookCorpus](https://yknzhu.wixsite.com/mbweb), a dataset consisting of 11,038 unpublished books and [English Wikipedia](https://en.wikipedia.org/wiki/English_Wikipedia) (excluding lists, tables and headers). ## Training procedure ### Preprocessing The texts are tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are then of the form: ``` [CLS] Sentence A [SEP] Sentence B [SEP] ``` With probability 0.5, sentence A and sentence B correspond to two consecutive sentences in the original corpus and in the other cases, it's another random sentence in the corpus. Note that what is considered a sentence here is a consecutive span of text usually longer than a single sentence. The only constrain is that the result with the two "sentences" has a combined length of less than 512 tokens. The details of the masking procedure for each sentence are the following: - 15% of the tokens are masked. - In 80% of the cases, the masked tokens are replaced by `[MASK]`. - In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace. - In the 10% remaining cases, the masked tokens are left as is. ### Pretraining The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total) for one million steps with a batch size of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer used is Adam with a learning rate of 1e-4, \$\beta_{1} = 0.9\$ and \$\beta_{2} = 0.999\$, a weight decay of 0.01, learning rate warmup for 10,000 steps and linear decay of the learning rate after. ## Evaluation results When fine-tuned on downstream tasks, this model achieves the following results: Glue test results: | Task | MNLI-(m/mm) | QQP | QNLI | SST-2 | CoLA | STS-B | MRPC | RTE | Average | |:----:|:-----------:|:----:|:----:|:-----:|:----:|:-----:|:----:|:----:|:-------:| | | 84.6/83.4 | 71.2 | 90.5 | 93.5 | 52.1 | 85.8 | 88.9 | 66.4 | 79.6 | ### BibTeX entry and citation info ```bibtex @article{DBLP:journals/corr/abs-1810-04805, author = {Jacob Devlin and Ming{-}Wei Chang and Kenton Lee and Kristina Toutanova}, title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language Understanding}, journal = {CoRR}, volume = {abs/1810.04805}, year = {2018}, url = {http://arxiv.org/abs/1810.04805}, archivePrefix = {arXiv}, eprint = {1810.04805}, timestamp = {Tue, 30 Oct 2018 20:39:56 +0100}, biburl = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ``` <a href="https://huggingface.co/exbert/?model=bert-base-cased"> <img width="300px" src="https://cdn-media.huggingface.co/exbert/button.png"> </a>

{"datasets": ["bookcorpus", "wikipedia"], "language": "en", "license": "apache-2.0", "tags": ["exbert"]}

task

[ "QUESTION_ANSWERING" ]

44,619

Lots-of-LoRAs/Mistral-7B-Instruct-v0.2-4b-r16-task1121

Lots-of-LoRAs

null

[ "pytorch", "safetensors", "en", "arxiv:1910.09700", "arxiv:2407.00066", "base_model:mistralai/Mistral-7B-Instruct-v0.2", "base_model:finetune:mistralai/Mistral-7B-Instruct-v0.2", "license:mit", "region:us" ]

2025-01-03T18:30:17Z

2025-01-03T18:30:23+00:00

0

--- base_model: mistralai/Mistral-7B-Instruct-v0.2 language: en library_name: pytorch license: mit --- # Model Card for Mistral-7B-Instruct-v0.2-4b-r16-task1121  ## Model Details ### Model Description  LoRA trained on task1121_alt_ja_khm_translation - **Developed by:** bruel - **Funded by [optional]:** [More Information Needed] - **Shared by [optional]:** [More Information Needed] - **Model type:** LoRA - **Language(s) (NLP):** en - **License:** mit - **Finetuned from model [optional]:** mistralai/Mistral-7B-Instruct-v0.2 ### Model Sources [optional]  - **Repository:** https://github.com/bruel-gabrielsson - **Paper [optional]:** "Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead" (2024), Rickard Brüel Gabrielsson, Jiacheng Zhu, Onkar Bhardwaj, Leshem Choshen, Kristjan Greenewald, Mikhail Yurochkin and Justin Solomon - **Demo [optional]:** [More Information Needed] ## Uses  ### Direct Use  [More Information Needed] ### Downstream Use [optional]  [More Information Needed] ### Out-of-Scope Use  [More Information Needed] ## Bias, Risks, and Limitations  [More Information Needed] ### Recommendations  Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data  https://huggingface.co/datasets/Lots-of-LoRAs/task1121_alt_ja_khm_translation sourced from https://github.com/allenai/natural-instructions ### Training Procedure  #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - **Training regime:** [More Information Needed]  #### Speeds, Sizes, Times [optional]  [More Information Needed] ## Evaluation  ### Testing Data, Factors & Metrics #### Testing Data  [More Information Needed] #### Factors  [More Information Needed] #### Metrics  [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional]  [More Information Needed] ## Environmental Impact  Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - **Hardware Type:** [More Information Needed] - **Hours used:** [More Information Needed] - **Cloud Provider:** [More Information Needed] - **Compute Region:** [More Information Needed] - **Carbon Emitted:** [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional]  **BibTeX:** @misc{brüelgabrielsson2024compressserveservingthousands, title={Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead}, author={Rickard Brüel-Gabrielsson and Jiacheng Zhu and Onkar Bhardwaj and Leshem Choshen and Kristjan Greenewald and Mikhail Yurochkin and Justin Solomon}, year={2024}, eprint={2407.00066}, archivePrefix={arXiv}, primaryClass={cs.DC}, url={https://arxiv.org/abs/2407.00066}, } **APA:** [More Information Needed] ## Glossary [optional]  [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]

null

Non_BioNLP

# Model Card for Mistral-7B-Instruct-v0.2-4b-r16-task1121  ## Model Details ### Model Description  LoRA trained on task1121_alt_ja_khm_translation - **Developed by:** bruel - **Funded by [optional]:** [More Information Needed] - **Shared by [optional]:** [More Information Needed] - **Model type:** LoRA - **Language(s) (NLP):** en - **License:** mit - **Finetuned from model [optional]:** mistralai/Mistral-7B-Instruct-v0.2 ### Model Sources [optional]  - **Repository:** https://github.com/bruel-gabrielsson - **Paper [optional]:** "Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead" (2024), Rickard Brüel Gabrielsson, Jiacheng Zhu, Onkar Bhardwaj, Leshem Choshen, Kristjan Greenewald, Mikhail Yurochkin and Justin Solomon - **Demo [optional]:** [More Information Needed] ## Uses  ### Direct Use  [More Information Needed] ### Downstream Use [optional]  [More Information Needed] ### Out-of-Scope Use  [More Information Needed] ## Bias, Risks, and Limitations  [More Information Needed] ### Recommendations  Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data  https://huggingface.co/datasets/Lots-of-LoRAs/task1121_alt_ja_khm_translation sourced from https://github.com/allenai/natural-instructions ### Training Procedure  #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - **Training regime:** [More Information Needed]  #### Speeds, Sizes, Times [optional]  [More Information Needed] ## Evaluation  ### Testing Data, Factors & Metrics #### Testing Data  [More Information Needed] #### Factors  [More Information Needed] #### Metrics  [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional]  [More Information Needed] ## Environmental Impact  Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - **Hardware Type:** [More Information Needed] - **Hours used:** [More Information Needed] - **Cloud Provider:** [More Information Needed] - **Compute Region:** [More Information Needed] - **Carbon Emitted:** [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional]  **BibTeX:** @misc{brüelgabrielsson2024compressserveservingthousands, title={Compress then Serve: Serving Thousands of LoRA Adapters with Little Overhead}, author={Rickard Brüel-Gabrielsson and Jiacheng Zhu and Onkar Bhardwaj and Leshem Choshen and Kristjan Greenewald and Mikhail Yurochkin and Justin Solomon}, year={2024}, eprint={2407.00066}, archivePrefix={arXiv}, primaryClass={cs.DC}, url={https://arxiv.org/abs/2407.00066}, } **APA:** [More Information Needed] ## Glossary [optional]  [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]

{"base_model": "mistralai/Mistral-7B-Instruct-v0.2", "language": "en", "library_name": "pytorch", "license": "mit"}

task

[ "TRANSLATION" ]

44,620

dz5035/bicleaner-ai-full-large-de-xx

dz5035

null

[ "transformers", "tf", "xlm-roberta", "bicleaner-ai", "ar", "bn", "da", "de", "en", "es", "fa", "hi", "ja", "ko", "ru", "ta", "th", "tr", "uk", "vi", "zh", "multilingual", "license:cc-by-sa-4.0", "endpoints_compatible", "region:us" ]

2024-06-03T13:20:12Z

2024-06-03T13:27:34+00:00

6

0

--- language: - ar - bn - da - de - en - es - fa - hi - ja - ko - ru - ta - th - tr - uk - vi - zh - multilingual license: cc-by-sa-4.0 tags: - bicleaner-ai tasks: - text-classification --- # Bicleaner AI multilingual full large model for de-xx Bicleaner AI is a tool that aims at detecting noisy sentence pairs in a parallel corpus. It indicates the likelihood of a pair of sentences being mutual translations (with a value near to 1) or not (with a value near to 0). Sentence pairs considered very noisy are scored with 0. Find out at our repository for further instructions on how to use it: https://github.com/bitextor/bicleaner-ai

null

Non_BioNLP

# Bicleaner AI multilingual full large model for de-xx Bicleaner AI is a tool that aims at detecting noisy sentence pairs in a parallel corpus. It indicates the likelihood of a pair of sentences being mutual translations (with a value near to 1) or not (with a value near to 0). Sentence pairs considered very noisy are scored with 0. Find out at our repository for further instructions on how to use it: https://github.com/bitextor/bicleaner-ai

{"language": ["ar", "bn", "da", "de", "en", "es", "fa", "hi", "ja", "ko", "ru", "ta", "th", "tr", "uk", "vi", "zh", "multilingual"], "license": "cc-by-sa-4.0", "tags": ["bicleaner-ai"], "tasks": ["text-classification"]}

task

[ "TRANSLATION" ]

44,621

lujain5666/finetune-gemma-7b

lujain5666

sentence-similarity

[ "sentence-transformers", "safetensors", "bert", "feature-extraction", "sentence-similarity", "transformers", "unsloth", "en", "dataset:s2orc", "dataset:flax-sentence-embeddings/stackexchange_xml", "dataset:ms_marco", "dataset:gooaq", "dataset:yahoo_answers_topics", "dataset:code_search_net", "dataset:search_qa", "dataset:eli5", "dataset:snli", "dataset:multi_nli", "dataset:wikihow", "dataset:natural_questions", "dataset:trivia_qa", "dataset:embedding-data/sentence-compression", "dataset:embedding-data/flickr30k-captions", "dataset:embedding-data/altlex", "dataset:embedding-data/simple-wiki", "dataset:embedding-data/QQP", "dataset:embedding-data/SPECTER", "dataset:embedding-data/PAQ_pairs", "dataset:embedding-data/WikiAnswers", "arxiv:1904.06472", "arxiv:2102.07033", "arxiv:2104.08727", "arxiv:1704.05179", "arxiv:1810.09305", "license:apache-2.0", "autotrain_compatible", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

2024-08-21T18:32:45Z

2024-08-22T14:49:25+00:00

9

0

--- datasets: - s2orc - flax-sentence-embeddings/stackexchange_xml - ms_marco - gooaq - yahoo_answers_topics - code_search_net - search_qa - eli5 - snli - multi_nli - wikihow - natural_questions - trivia_qa - embedding-data/sentence-compression - embedding-data/flickr30k-captions - embedding-data/altlex - embedding-data/simple-wiki - embedding-data/QQP - embedding-data/SPECTER - embedding-data/PAQ_pairs - embedding-data/WikiAnswers language: en library_name: sentence-transformers license: apache-2.0 pipeline_tag: sentence-similarity tags: - sentence-transformers - feature-extraction - sentence-similarity - transformers - unsloth --- # all-MiniLM-L6-v2 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer sentences = ["This is an example sentence", "Each sentence is converted"] model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2') embeddings = model.encode(sentences) print(embeddings) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch import torch.nn.functional as F #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) # Sentences we want sentence embeddings for sentences = ['This is an example sentence', 'Each sentence is converted'] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') model = AutoModel.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') # Tokenize sentences encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(**encoded_input) # Perform pooling sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask']) # Normalize embeddings sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1) print("Sentence embeddings:") print(sentence_embeddings) ``` ## Evaluation Results For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/all-MiniLM-L6-v2) ------ ## Background The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model and fine-tuned in on a 1B sentence pairs dataset. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset. We developed this model during the [Community week using JAX/Flax for NLP & CV](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104), organized by Hugging Face. We developed this model as part of the project: [Train the Best Sentence Embedding Model Ever with 1B Training Pairs](https://discuss.huggingface.co/t/train-the-best-sentence-embedding-model-ever-with-1b-training-pairs/7354). We benefited from efficient hardware infrastructure to run the project: 7 TPUs v3-8, as well as intervention from Googles Flax, JAX, and Cloud team member about efficient deep learning frameworks. ## Intended uses Our model is intended to be used as a sentence and short paragraph encoder. Given an input text, it outputs a vector which captures the semantic information. The sentence vector may be used for information retrieval, clustering or sentence similarity tasks. By default, input text longer than 256 word pieces is truncated. ## Training procedure ### Pre-training We use the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model. Please refer to the model card for more detailed information about the pre-training procedure. ### Fine-tuning We fine-tune the model using a contrastive objective. Formally, we compute the cosine similarity from each possible sentence pairs from the batch. We then apply the cross entropy loss by comparing with true pairs. #### Hyper parameters We trained our model on a TPU v3-8. We train the model during 100k steps using a batch size of 1024 (128 per TPU core). We use a learning rate warm up of 500. The sequence length was limited to 128 tokens. We used the AdamW optimizer with a 2e-5 learning rate. The full training script is accessible in this current repository: `train_script.py`. #### Training data We use the concatenation from multiple datasets to fine-tune our model. The total number of sentence pairs is above 1 billion sentences. We sampled each dataset given a weighted probability which configuration is detailed in the `data_config.json` file. | Dataset | Paper | Number of training tuples | |--------------------------------------------------------|:----------------------------------------:|:--------------------------:| | [Reddit comments (2015-2018)](https://github.com/PolyAI-LDN/conversational-datasets/tree/master/reddit) | [paper](https://arxiv.org/abs/1904.06472) | 726,484,430 | | [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Abstracts) | [paper](https://aclanthology.org/2020.acl-main.447/) | 116,288,806 | | [WikiAnswers](https://github.com/afader/oqa#wikianswers-corpus) Duplicate question pairs | [paper](https://doi.org/10.1145/2623330.2623677) | 77,427,422 | | [PAQ](https://github.com/facebookresearch/PAQ) (Question, Answer) pairs | [paper](https://arxiv.org/abs/2102.07033) | 64,371,441 | | [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Titles) | [paper](https://aclanthology.org/2020.acl-main.447/) | 52,603,982 | | [S2ORC](https://github.com/allenai/s2orc) (Title, Abstract) | [paper](https://aclanthology.org/2020.acl-main.447/) | 41,769,185 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Body) pairs | - | 25,316,456 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title+Body, Answer) pairs | - | 21,396,559 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Answer) pairs | - | 21,396,559 | | [MS MARCO](https://microsoft.github.io/msmarco/) triplets | [paper](https://doi.org/10.1145/3404835.3462804) | 9,144,553 | | [GOOAQ: Open Question Answering with Diverse Answer Types](https://github.com/allenai/gooaq) | [paper](https://arxiv.org/pdf/2104.08727.pdf) | 3,012,496 | | [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Answer) | [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) | 1,198,260 | | [Code Search](https://huggingface.co/datasets/code_search_net) | - | 1,151,414 | | [COCO](https://cocodataset.org/#home) Image captions | [paper](https://link.springer.com/chapter/10.1007%2F978-3-319-10602-1_48) | 828,395| | [SPECTER](https://github.com/allenai/specter) citation triplets | [paper](https://doi.org/10.18653/v1/2020.acl-main.207) | 684,100 | | [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Question, Answer) | [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) | 681,164 | | [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Question) | [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) | 659,896 | | [SearchQA](https://huggingface.co/datasets/search_qa) | [paper](https://arxiv.org/abs/1704.05179) | 582,261 | | [Eli5](https://huggingface.co/datasets/eli5) | [paper](https://doi.org/10.18653/v1/p19-1346) | 325,475 | | [Flickr 30k](https://shannon.cs.illinois.edu/DenotationGraph/) | [paper](https://transacl.org/ojs/index.php/tacl/article/view/229/33) | 317,695 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles) | | 304,525 | | AllNLI ([SNLI](https://nlp.stanford.edu/projects/snli/) and [MultiNLI](https://cims.nyu.edu/~sbowman/multinli/) | [paper SNLI](https://doi.org/10.18653/v1/d15-1075), [paper MultiNLI](https://doi.org/10.18653/v1/n18-1101) | 277,230 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (bodies) | | 250,519 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles+bodies) | | 250,460 | | [Sentence Compression](https://github.com/google-research-datasets/sentence-compression) | [paper](https://www.aclweb.org/anthology/D13-1155/) | 180,000 | | [Wikihow](https://github.com/pvl/wikihow_pairs_dataset) | [paper](https://arxiv.org/abs/1810.09305) | 128,542 | | [Altlex](https://github.com/chridey/altlex/) | [paper](https://aclanthology.org/P16-1135.pdf) | 112,696 | | [Quora Question Triplets](https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs) | - | 103,663 | | [Simple Wikipedia](https://cs.pomona.edu/~dkauchak/simplification/) | [paper](https://www.aclweb.org/anthology/P11-2117/) | 102,225 | | [Natural Questions (NQ)](https://ai.google.com/research/NaturalQuestions) | [paper](https://transacl.org/ojs/index.php/tacl/article/view/1455) | 100,231 | | [SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) | [paper](https://aclanthology.org/P18-2124.pdf) | 87,599 | | [TriviaQA](https://huggingface.co/datasets/trivia_qa) | - | 73,346 | | **Total** | | **1,170,060,424** |

null

Non_BioNLP

# all-MiniLM-L6-v2 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer sentences = ["This is an example sentence", "Each sentence is converted"] model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2') embeddings = model.encode(sentences) print(embeddings) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch import torch.nn.functional as F #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) # Sentences we want sentence embeddings for sentences = ['This is an example sentence', 'Each sentence is converted'] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') model = AutoModel.from_pretrained('sentence-transformers/all-MiniLM-L6-v2') # Tokenize sentences encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(**encoded_input) # Perform pooling sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask']) # Normalize embeddings sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1) print("Sentence embeddings:") print(sentence_embeddings) ``` ## Evaluation Results For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/all-MiniLM-L6-v2) ------ ## Background The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model and fine-tuned in on a 1B sentence pairs dataset. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset. We developed this model during the [Community week using JAX/Flax for NLP & CV](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104), organized by Hugging Face. We developed this model as part of the project: [Train the Best Sentence Embedding Model Ever with 1B Training Pairs](https://discuss.huggingface.co/t/train-the-best-sentence-embedding-model-ever-with-1b-training-pairs/7354). We benefited from efficient hardware infrastructure to run the project: 7 TPUs v3-8, as well as intervention from Googles Flax, JAX, and Cloud team member about efficient deep learning frameworks. ## Intended uses Our model is intended to be used as a sentence and short paragraph encoder. Given an input text, it outputs a vector which captures the semantic information. The sentence vector may be used for information retrieval, clustering or sentence similarity tasks. By default, input text longer than 256 word pieces is truncated. ## Training procedure ### Pre-training We use the pretrained [`nreimers/MiniLM-L6-H384-uncased`](https://huggingface.co/nreimers/MiniLM-L6-H384-uncased) model. Please refer to the model card for more detailed information about the pre-training procedure. ### Fine-tuning We fine-tune the model using a contrastive objective. Formally, we compute the cosine similarity from each possible sentence pairs from the batch. We then apply the cross entropy loss by comparing with true pairs. #### Hyper parameters We trained our model on a TPU v3-8. We train the model during 100k steps using a batch size of 1024 (128 per TPU core). We use a learning rate warm up of 500. The sequence length was limited to 128 tokens. We used the AdamW optimizer with a 2e-5 learning rate. The full training script is accessible in this current repository: `train_script.py`. #### Training data We use the concatenation from multiple datasets to fine-tune our model. The total number of sentence pairs is above 1 billion sentences. We sampled each dataset given a weighted probability which configuration is detailed in the `data_config.json` file. | Dataset | Paper | Number of training tuples | |--------------------------------------------------------|:----------------------------------------:|:--------------------------:| | [Reddit comments (2015-2018)](https://github.com/PolyAI-LDN/conversational-datasets/tree/master/reddit) | [paper](https://arxiv.org/abs/1904.06472) | 726,484,430 | | [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Abstracts) | [paper](https://aclanthology.org/2020.acl-main.447/) | 116,288,806 | | [WikiAnswers](https://github.com/afader/oqa#wikianswers-corpus) Duplicate question pairs | [paper](https://doi.org/10.1145/2623330.2623677) | 77,427,422 | | [PAQ](https://github.com/facebookresearch/PAQ) (Question, Answer) pairs | [paper](https://arxiv.org/abs/2102.07033) | 64,371,441 | | [S2ORC](https://github.com/allenai/s2orc) Citation pairs (Titles) | [paper](https://aclanthology.org/2020.acl-main.447/) | 52,603,982 | | [S2ORC](https://github.com/allenai/s2orc) (Title, Abstract) | [paper](https://aclanthology.org/2020.acl-main.447/) | 41,769,185 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Body) pairs | - | 25,316,456 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title+Body, Answer) pairs | - | 21,396,559 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Answer) pairs | - | 21,396,559 | | [MS MARCO](https://microsoft.github.io/msmarco/) triplets | [paper](https://doi.org/10.1145/3404835.3462804) | 9,144,553 | | [GOOAQ: Open Question Answering with Diverse Answer Types](https://github.com/allenai/gooaq) | [paper](https://arxiv.org/pdf/2104.08727.pdf) | 3,012,496 | | [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Answer) | [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) | 1,198,260 | | [Code Search](https://huggingface.co/datasets/code_search_net) | - | 1,151,414 | | [COCO](https://cocodataset.org/#home) Image captions | [paper](https://link.springer.com/chapter/10.1007%2F978-3-319-10602-1_48) | 828,395| | [SPECTER](https://github.com/allenai/specter) citation triplets | [paper](https://doi.org/10.18653/v1/2020.acl-main.207) | 684,100 | | [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Question, Answer) | [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) | 681,164 | | [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Question) | [paper](https://proceedings.neurips.cc/paper/2015/hash/250cf8b51c773f3f8dc8b4be867a9a02-Abstract.html) | 659,896 | | [SearchQA](https://huggingface.co/datasets/search_qa) | [paper](https://arxiv.org/abs/1704.05179) | 582,261 | | [Eli5](https://huggingface.co/datasets/eli5) | [paper](https://doi.org/10.18653/v1/p19-1346) | 325,475 | | [Flickr 30k](https://shannon.cs.illinois.edu/DenotationGraph/) | [paper](https://transacl.org/ojs/index.php/tacl/article/view/229/33) | 317,695 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles) | | 304,525 | | AllNLI ([SNLI](https://nlp.stanford.edu/projects/snli/) and [MultiNLI](https://cims.nyu.edu/~sbowman/multinli/) | [paper SNLI](https://doi.org/10.18653/v1/d15-1075), [paper MultiNLI](https://doi.org/10.18653/v1/n18-1101) | 277,230 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (bodies) | | 250,519 | | [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions (titles+bodies) | | 250,460 | | [Sentence Compression](https://github.com/google-research-datasets/sentence-compression) | [paper](https://www.aclweb.org/anthology/D13-1155/) | 180,000 | | [Wikihow](https://github.com/pvl/wikihow_pairs_dataset) | [paper](https://arxiv.org/abs/1810.09305) | 128,542 | | [Altlex](https://github.com/chridey/altlex/) | [paper](https://aclanthology.org/P16-1135.pdf) | 112,696 | | [Quora Question Triplets](https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs) | - | 103,663 | | [Simple Wikipedia](https://cs.pomona.edu/~dkauchak/simplification/) | [paper](https://www.aclweb.org/anthology/P11-2117/) | 102,225 | | [Natural Questions (NQ)](https://ai.google.com/research/NaturalQuestions) | [paper](https://transacl.org/ojs/index.php/tacl/article/view/1455) | 100,231 | | [SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) | [paper](https://aclanthology.org/P18-2124.pdf) | 87,599 | | [TriviaQA](https://huggingface.co/datasets/trivia_qa) | - | 73,346 | | **Total** | | **1,170,060,424** |

{"datasets": ["s2orc", "flax-sentence-embeddings/stackexchange_xml", "ms_marco", "gooaq", "yahoo_answers_topics", "code_search_net", "search_qa", "eli5", "snli", "multi_nli", "wikihow", "natural_questions", "trivia_qa", "embedding-data/sentence-compression", "embedding-data/flickr30k-captions", "embedding-data/altlex", "embedding-data/simple-wiki", "embedding-data/QQP", "embedding-data/SPECTER", "embedding-data/PAQ_pairs", "embedding-data/WikiAnswers"], "language": "en", "library_name": "sentence-transformers", "license": "apache-2.0", "pipeline_tag": "sentence-similarity", "tags": ["sentence-transformers", "feature-extraction", "sentence-similarity", "transformers", "unsloth"]}

task

[ "QUESTION_ANSWERING" ]

44,622

amitesh11/bart-finance-pegasus

amitesh11

summarization

[ "transformers", "pytorch", "bart", "text2text-generation", "summarization", "en", "license:ms-pl", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-10-17T11:14:14Z

2023-10-17T11:32:10+00:00

111

1

--- language: - en license: ms-pl pipeline_tag: summarization ---

null

Non_BioNLP

{"language": ["en"], "license": "ms-pl", "pipeline_tag": "summarization"}

task

[ "SUMMARIZATION" ]

44,623

UBC-NLP/AraT5-base-title-generation

UBC-NLP

text2text-generation

[ "transformers", "pytorch", "tf", "t5", "text2text-generation", "Arabic T5", "MSA", "Twitter", "Arabic Dialect", "Arabic Machine Translation", "Arabic Text Summarization", "Arabic News Title and Question Generation", "Arabic Paraphrasing and Transliteration", "Arabic Code-Switched Translation", "ar", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:05Z

2022-05-26T18:29:45+00:00

50

12

--- language: - ar tags: - Arabic T5 - MSA - Twitter - Arabic Dialect - Arabic Machine Translation - Arabic Text Summarization - Arabic News Title and Question Generation - Arabic Paraphrasing and Transliteration - Arabic Code-Switched Translation --- # AraT5-base-title-generation # AraT5: Text-to-Text Transformers for Arabic Language Generation <img src="https://huggingface.co/UBC-NLP/AraT5-base/resolve/main/AraT5_CR_new.png" alt="AraT5" width="45%" height="35%" align="right"/> This is the repository accompanying our paper [AraT5: Text-to-Text Transformers for Arabic Language Understanding and Generation](https://aclanthology.org/2022.acl-long.47/). In this is the repository we Introduce **AraT5MSA**, **AraT5Tweet**, and **AraT5**: three powerful Arabic-specific text-to-text Transformer based models; --- # How to use AraT5 models Below is an example for fine-tuning **AraT5-base** for News Title Generation on the Aranews dataset ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("UBC-NLP/AraT5-base-title-generation") model = AutoModelForSeq2SeqLM.from_pretrained("UBC-NLP/AraT5-base-title-generation") Document = "تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة ." encoding = tokenizer.encode_plus(Document,pad_to_max_length=True, return_tensors="pt") input_ids, attention_masks = encoding["input_ids"], encoding["attention_mask"] outputs = model.generate( input_ids=input_ids, attention_mask=attention_masks, max_length=256, do_sample=True, top_k=120, top_p=0.95, early_stopping=True, num_return_sequences=5 ) for id, output in enumerate(outputs): title = tokenizer.decode(output, skip_special_tokens=True,clean_up_tokenization_spaces=True) print("title#"+str(id), title) ``` **The input news document** <div style="white-space : pre-wrap !important;word-break: break-word; direction:rtl; text-align: right"> تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة . </div> **The generated titles** ``` title#0 غرفة الشرقية تختتم المرحلة الثانية من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة title#1 غرفة الشرقية تختتم الثاني من مبادرة تأهيل وتأهيل أبناء وبناتنا title#2 سعود بن نايف يختتم ثانى مبادراتها لتأهيل وتدريب أبناء وبنات المملكة title#3 أمير الشرقية يرعى اختتام برنامج برنامج تدريب أبناء وبنات المملكة title#4 سعود بن نايف يرعى اختتام مبادرة تأهيل وتدريب أبناء وبنات المملكة ``` # AraT5 Models Checkpoints AraT5 Pytorch and TensorFlow checkpoints are available on the Huggingface website for direct download and use ```exclusively for research```. ```For commercial use, please contact the authors via email @ (muhammad.mageed[at]ubc[dot]ca).``` | **Model** | **Link** | |---------|:------------------:| | **AraT5-base** | [https://huggingface.co/UBC-NLP/AraT5-base](https://huggingface.co/UBC-NLP/AraT5-base) | | **AraT5-msa-base** | [https://huggingface.co/UBC-NLP/AraT5-msa-base](https://huggingface.co/UBC-NLP/AraT5-msa-base) | | **AraT5-tweet-base** | [https://huggingface.co/UBC-NLP/AraT5-tweet-base](https://huggingface.co/UBC-NLP/AraT5-tweet-base) | | **AraT5-msa-small** | [https://huggingface.co/UBC-NLP/AraT5-msa-small](https://huggingface.co/UBC-NLP/AraT5-msa-small) | | **AraT5-tweet-small**| [https://huggingface.co/UBC-NLP/AraT5-tweet-small](https://huggingface.co/UBC-NLP/AraT5-tweet-small) | # BibTex If you use our models (Arat5-base, Arat5-msa-base, Arat5-tweet-base, Arat5-msa-small, or Arat5-tweet-small ) for your scientific publication, or if you find the resources in this repository useful, please cite our paper as follows (to be updated): ```bibtex @inproceedings{nagoudi-etal-2022-arat5, title = "{A}ra{T}5: Text-to-Text Transformers for {A}rabic Language Generation", author = "Nagoudi, El Moatez Billah and Elmadany, AbdelRahim and Abdul-Mageed, Muhammad", booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = may, year = "2022", address = "Dublin, Ireland", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2022.acl-long.47", pages = "628--647", abstract = "Transfer learning with a unified Transformer framework (T5) that converts all language problems into a text-to-text format was recently proposed as a simple and effective transfer learning approach. Although a multilingual version of the T5 model (mT5) was also introduced, it is not clear how well it can fare on non-English tasks involving diverse data. To investigate this question, we apply mT5 on a language with a wide variety of dialects{--}Arabic. For evaluation, we introduce a novel benchmark for ARabic language GENeration (ARGEN), covering seven important tasks. For model comparison, we pre-train three powerful Arabic T5-style models and evaluate them on ARGEN. Although pre-trained with {\textasciitilde}49 less data, our new models perform significantly better than mT5 on all ARGEN tasks (in 52 out of 59 test sets) and set several new SOTAs. Our models also establish new SOTA on the recently-proposed, large Arabic language understanding evaluation benchmark ARLUE (Abdul-Mageed et al., 2021). Our new models are publicly available. We also link to ARGEN datasets through our repository: https://github.com/UBC-NLP/araT5.", } ``` ## Acknowledgments We gratefully acknowledge support from the Natural Sciences and Engineering Research Council of Canada, the Social Sciences and Humanities Research Council of Canada, Canadian Foundation for Innovation, [ComputeCanada](www.computecanada.ca) and [UBC ARC-Sockeye](https://doi.org/10.14288/SOCKEYE). We also thank the [Google TensorFlow Research Cloud (TFRC)](https://www.tensorflow.org/tfrc) program for providing us with free TPU access.

null

Non_BioNLP

# AraT5-base-title-generation # AraT5: Text-to-Text Transformers for Arabic Language Generation <img src="https://huggingface.co/UBC-NLP/AraT5-base/resolve/main/AraT5_CR_new.png" alt="AraT5" width="45%" height="35%" align="right"/> This is the repository accompanying our paper [AraT5: Text-to-Text Transformers for Arabic Language Understanding and Generation](https://aclanthology.org/2022.acl-long.47/). In this is the repository we Introduce **AraT5MSA**, **AraT5Tweet**, and **AraT5**: three powerful Arabic-specific text-to-text Transformer based models; --- # How to use AraT5 models Below is an example for fine-tuning **AraT5-base** for News Title Generation on the Aranews dataset ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("UBC-NLP/AraT5-base-title-generation") model = AutoModelForSeq2SeqLM.from_pretrained("UBC-NLP/AraT5-base-title-generation") Document = "تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة ." encoding = tokenizer.encode_plus(Document,pad_to_max_length=True, return_tensors="pt") input_ids, attention_masks = encoding["input_ids"], encoding["attention_mask"] outputs = model.generate( input_ids=input_ids, attention_mask=attention_masks, max_length=256, do_sample=True, top_k=120, top_p=0.95, early_stopping=True, num_return_sequences=5 ) for id, output in enumerate(outputs): title = tokenizer.decode(output, skip_special_tokens=True,clean_up_tokenization_spaces=True) print("title#"+str(id), title) ``` **The input news document** <div style="white-space : pre-wrap !important;word-break: break-word; direction:rtl; text-align: right"> تحت رعاية صاحب السمو الملكي الأمير سعود بن نايف بن عبدالعزيز أمير المنطقة الشرقية اختتمت غرفة الشرقية مؤخرا، الثاني من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة ضمن مبادرتها المجانية للعام 2019 حيث قدمت 6 برامج تدريبية نوعية. وثمن رئيس مجلس إدارة الغرفة، عبدالحكيم العمار الخالدي، رعاية سمو أمير المنطقة الشرقية للمبادرة، مؤكدا أن دعم سموه لجميع أنشطة . </div> **The generated titles** ``` title#0 غرفة الشرقية تختتم المرحلة الثانية من مبادرتها لتأهيل وتدريب أبناء وبنات المملكة title#1 غرفة الشرقية تختتم الثاني من مبادرة تأهيل وتأهيل أبناء وبناتنا title#2 سعود بن نايف يختتم ثانى مبادراتها لتأهيل وتدريب أبناء وبنات المملكة title#3 أمير الشرقية يرعى اختتام برنامج برنامج تدريب أبناء وبنات المملكة title#4 سعود بن نايف يرعى اختتام مبادرة تأهيل وتدريب أبناء وبنات المملكة ``` # AraT5 Models Checkpoints AraT5 Pytorch and TensorFlow checkpoints are available on the Huggingface website for direct download and use ```exclusively for research```. ```For commercial use, please contact the authors via email @ (muhammad.mageed[at]ubc[dot]ca).``` | **Model** | **Link** | |---------|:------------------:| | **AraT5-base** | [https://huggingface.co/UBC-NLP/AraT5-base](https://huggingface.co/UBC-NLP/AraT5-base) | | **AraT5-msa-base** | [https://huggingface.co/UBC-NLP/AraT5-msa-base](https://huggingface.co/UBC-NLP/AraT5-msa-base) | | **AraT5-tweet-base** | [https://huggingface.co/UBC-NLP/AraT5-tweet-base](https://huggingface.co/UBC-NLP/AraT5-tweet-base) | | **AraT5-msa-small** | [https://huggingface.co/UBC-NLP/AraT5-msa-small](https://huggingface.co/UBC-NLP/AraT5-msa-small) | | **AraT5-tweet-small**| [https://huggingface.co/UBC-NLP/AraT5-tweet-small](https://huggingface.co/UBC-NLP/AraT5-tweet-small) | # BibTex If you use our models (Arat5-base, Arat5-msa-base, Arat5-tweet-base, Arat5-msa-small, or Arat5-tweet-small ) for your scientific publication, or if you find the resources in this repository useful, please cite our paper as follows (to be updated): ```bibtex @inproceedings{nagoudi-etal-2022-arat5, title = "{A}ra{T}5: Text-to-Text Transformers for {A}rabic Language Generation", author = "Nagoudi, El Moatez Billah and Elmadany, AbdelRahim and Abdul-Mageed, Muhammad", booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = may, year = "2022", address = "Dublin, Ireland", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2022.acl-long.47", pages = "628--647", abstract = "Transfer learning with a unified Transformer framework (T5) that converts all language problems into a text-to-text format was recently proposed as a simple and effective transfer learning approach. Although a multilingual version of the T5 model (mT5) was also introduced, it is not clear how well it can fare on non-English tasks involving diverse data. To investigate this question, we apply mT5 on a language with a wide variety of dialects{--}Arabic. For evaluation, we introduce a novel benchmark for ARabic language GENeration (ARGEN), covering seven important tasks. For model comparison, we pre-train three powerful Arabic T5-style models and evaluate them on ARGEN. Although pre-trained with {\textasciitilde}49 less data, our new models perform significantly better than mT5 on all ARGEN tasks (in 52 out of 59 test sets) and set several new SOTAs. Our models also establish new SOTA on the recently-proposed, large Arabic language understanding evaluation benchmark ARLUE (Abdul-Mageed et al., 2021). Our new models are publicly available. We also link to ARGEN datasets through our repository: https://github.com/UBC-NLP/araT5.", } ``` ## Acknowledgments We gratefully acknowledge support from the Natural Sciences and Engineering Research Council of Canada, the Social Sciences and Humanities Research Council of Canada, Canadian Foundation for Innovation, [ComputeCanada](www.computecanada.ca) and [UBC ARC-Sockeye](https://doi.org/10.14288/SOCKEYE). We also thank the [Google TensorFlow Research Cloud (TFRC)](https://www.tensorflow.org/tfrc) program for providing us with free TPU access.

{"language": ["ar"], "tags": ["Arabic T5", "MSA", "Twitter", "Arabic Dialect", "Arabic Machine Translation", "Arabic Text Summarization", "Arabic News Title and Question Generation", "Arabic Paraphrasing and Transliteration", "Arabic Code-Switched Translation"]}

task

[ "TRANSLATION", "SUMMARIZATION", "PARAPHRASING" ]

44,624

Dhahlan2000/Simple_Translation-model-for-GPT-v13

Dhahlan2000

text2text-generation

[ "transformers", "tensorboard", "safetensors", "t5", "text2text-generation", "generated_from_trainer", "base_model:Dhahlan2000/Simple_Translation-model-for-GPT-v12", "base_model:finetune:Dhahlan2000/Simple_Translation-model-for-GPT-v12", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

2024-06-09T07:39:40Z

2024-06-09T07:40:02+00:00

4

0

--- base_model: Dhahlan2000/Simple_Translation-model-for-GPT-v12 metrics: - bleu tags: - generated_from_trainer model-index: - name: Simple_Translation-model-for-GPT-v13 results: [] ---  # Simple_Translation-model-for-GPT-v13 This model is a fine-tuned version of [Dhahlan2000/Simple_Translation-model-for-GPT-v12](https://huggingface.co/Dhahlan2000/Simple_Translation-model-for-GPT-v12) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2668 - Bleu: 17.0937 - Gen Len: 18.5927 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:| | 0.3914 | 1.0 | 9282 | 0.2714 | 16.9453 | 18.5893 | | 0.3774 | 2.0 | 18564 | 0.2668 | 17.0937 | 18.5927 | ### Framework versions - Transformers 4.41.2 - Pytorch 2.3.0+cu121 - Datasets 2.19.2 - Tokenizers 0.19.1

null

Non_BioNLP

# Simple_Translation-model-for-GPT-v13 This model is a fine-tuned version of [Dhahlan2000/Simple_Translation-model-for-GPT-v12](https://huggingface.co/Dhahlan2000/Simple_Translation-model-for-GPT-v12) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2668 - Bleu: 17.0937 - Gen Len: 18.5927 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:| | 0.3914 | 1.0 | 9282 | 0.2714 | 16.9453 | 18.5893 | | 0.3774 | 2.0 | 18564 | 0.2668 | 17.0937 | 18.5927 | ### Framework versions - Transformers 4.41.2 - Pytorch 2.3.0+cu121 - Datasets 2.19.2 - Tokenizers 0.19.1

{"base_model": "Dhahlan2000/Simple_Translation-model-for-GPT-v12", "metrics": ["bleu"], "tags": ["generated_from_trainer"], "model-index": [{"name": "Simple_Translation-model-for-GPT-v13", "results": []}]}

task

[ "TRANSLATION" ]

44,625

cross-encoder/stsb-roberta-base

cross-encoder

text-classification

[ "transformers", "pytorch", "jax", "safetensors", "roberta", "text-classification", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:05Z

2025-04-11T08:18:13+00:00

202,254

4

--- base_model: - FacebookAI/roberta-base datasets: - sentence-transformers/stsb language: - en library_name: sentence-transformers license: apache-2.0 pipeline_tag: text-ranking tags: - transformers --- # Cross-Encoder for Semantic Textual Similarity This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class. ## Training Data This model was trained on the [STS benchmark dataset](http://ixa2.si.ehu.eus/stswiki/index.php/STSbenchmark). The model will predict a score between 0 and 1 how for the semantic similarity of two sentences. ## Usage and Performance Pre-trained models can be used like this: ```python from sentence_transformers import CrossEncoder model = CrossEncoder('cross-encoder/stsb-roberta-base') scores = model.predict([('Sentence 1', 'Sentence 2'), ('Sentence 3', 'Sentence 4')]) ``` The model will predict scores for the pairs `('Sentence 1', 'Sentence 2')` and `('Sentence 3', 'Sentence 4')`. You can use this model also without sentence_transformers and by just using Transformers ``AutoModel`` class

null

Non_BioNLP

# Cross-Encoder for Semantic Textual Similarity This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class. ## Training Data This model was trained on the [STS benchmark dataset](http://ixa2.si.ehu.eus/stswiki/index.php/STSbenchmark). The model will predict a score between 0 and 1 how for the semantic similarity of two sentences. ## Usage and Performance Pre-trained models can be used like this: ```python from sentence_transformers import CrossEncoder model = CrossEncoder('cross-encoder/stsb-roberta-base') scores = model.predict([('Sentence 1', 'Sentence 2'), ('Sentence 3', 'Sentence 4')]) ``` The model will predict scores for the pairs `('Sentence 1', 'Sentence 2')` and `('Sentence 3', 'Sentence 4')`. You can use this model also without sentence_transformers and by just using Transformers ``AutoModel`` class

{"base_model": ["FacebookAI/roberta-base"], "datasets": ["sentence-transformers/stsb"], "language": ["en"], "library_name": "sentence-transformers", "license": "apache-2.0", "pipeline_tag": "text-ranking", "tags": ["transformers"]}

task

[ "SEMANTIC_SIMILARITY" ]

44,626

Helsinki-NLP/opus-mt-fi-guw

Helsinki-NLP

translation

[ "transformers", "pytorch", "tf", "marian", "text2text-generation", "translation", "fi", "guw", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2022-03-02T23:29:04Z

2023-08-16T11:34:37+00:00

48

0

--- license: apache-2.0 tags: - translation --- ### opus-mt-fi-guw * source languages: fi * target languages: guw * OPUS readme: [fi-guw](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fi-guw/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-08.zip](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.zip) * test set translations: [opus-2020-01-08.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.test.txt) * test set scores: [opus-2020-01-08.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.eval.txt) ## Benchmarks | testset | BLEU | chr-F | |-----------------------|-------|-------| | JW300.fi.guw | 32.4 | 0.527 |

null

Non_BioNLP

### opus-mt-fi-guw * source languages: fi * target languages: guw * OPUS readme: [fi-guw](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fi-guw/README.md) * dataset: opus * model: transformer-align * pre-processing: normalization + SentencePiece * download original weights: [opus-2020-01-08.zip](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.zip) * test set translations: [opus-2020-01-08.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.test.txt) * test set scores: [opus-2020-01-08.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fi-guw/opus-2020-01-08.eval.txt) ## Benchmarks | testset | BLEU | chr-F | |-----------------------|-------|-------| | JW300.fi.guw | 32.4 | 0.527 |

{"license": "apache-2.0", "tags": ["translation"]}

task

[ "TRANSLATION" ]

44,627

vishnupriyavr/flan-t5-movie-summary

vishnupriyavr

summarization

[ "peft", "art", "summarization", "en", "dataset:vishnupriyavr/wiki-movie-plots-with-summaries", "license:apache-2.0", "region:us" ]

2023-10-07T13:51:27Z

2023-10-07T20:34:48+00:00

2

0

--- datasets: - vishnupriyavr/wiki-movie-plots-with-summaries language: - en library_name: peft license: apache-2.0 metrics: - rouge pipeline_tag: summarization tags: - art ---

null

Non_BioNLP

{"datasets": ["vishnupriyavr/wiki-movie-plots-with-summaries"], "language": ["en"], "library_name": "peft", "license": "apache-2.0", "metrics": ["rouge"], "pipeline_tag": "summarization", "tags": ["art"]}

task

[ "SUMMARIZATION" ]

44,628

Luciferio/MiniLLM-finetuned

Luciferio

text-classification

[ "transformers", "pytorch", "safetensors", "bert", "text-classification", "generated_from_trainer", "dataset:emotion", "base_model:microsoft/MiniLM-L12-H384-uncased", "base_model:finetune:microsoft/MiniLM-L12-H384-uncased", "license:mit", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

2023-08-22T12:57:56Z

2023-08-22T14:39:58+00:00

14

0

--- base_model: microsoft/MiniLM-L12-H384-uncased datasets: - emotion license: mit metrics: - f1 tags: - generated_from_trainer model-index: - name: MiniLLM-finetuned results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion config: split split: validation args: split metrics: - type: f1 value: 0.922353805579638 name: F1 ---  # MiniLLM-finetuned This model is a fine-tuned version of [microsoft/MiniLM-L12-H384-uncased](https://huggingface.co/microsoft/MiniLM-L12-H384-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2932 - F1: 0.9224 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | F1 | |:-------------:|:-----:|:----:|:---------------:|:------:| | No log | 1.0 | 2000 | 0.4408 | 0.8888 | | No log | 2.0 | 4000 | 0.2932 | 0.9224 | ### Framework versions - Transformers 4.31.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

null

Non_BioNLP

# MiniLLM-finetuned This model is a fine-tuned version of [microsoft/MiniLM-L12-H384-uncased](https://huggingface.co/microsoft/MiniLM-L12-H384-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2932 - F1: 0.9224 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | F1 | |:-------------:|:-----:|:----:|:---------------:|:------:| | No log | 1.0 | 2000 | 0.4408 | 0.8888 | | No log | 2.0 | 4000 | 0.2932 | 0.9224 | ### Framework versions - Transformers 4.31.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

{"base_model": "microsoft/MiniLM-L12-H384-uncased", "datasets": ["emotion"], "license": "mit", "metrics": ["f1"], "tags": ["generated_from_trainer"], "model-index": [{"name": "MiniLLM-finetuned", "results": [{"task": {"type": "text-classification", "name": "Text Classification"}, "dataset": {"name": "emotion", "type": "emotion", "config": "split", "split": "validation", "args": "split"}, "metrics": [{"type": "f1", "value": 0.922353805579638, "name": "F1"}]}]}]}

task

[ "TEXT_CLASSIFICATION" ]

44,629

ccdv/lsg-bart-base-4096-wcep

ccdv

summarization

[ "transformers", "pytorch", "bart", "text2text-generation", "summarization", "custom_code", "en", "dataset:ccdv/WCEP-10", "arxiv:2210.15497", "autotrain_compatible", "region:us" ]

2022-05-25T11:09:11Z

2023-12-17T21:10:24+00:00

50

2

--- datasets: - ccdv/WCEP-10 language: - en metrics: - rouge tags: - summarization model-index: - name: ccdv/lsg-bart-base-4096-wcep results: [] ---  **Transformers >= 4.36.1**\ **This model relies on a custom modeling file, you need to add trust_remote_code=True**\ **See [\#13467](https://github.com/huggingface/transformers/pull/13467)** LSG ArXiv [paper](https://arxiv.org/abs/2210.15497). \ Github/conversion script is available at this [link](https://github.com/ccdv-ai/convert_checkpoint_to_lsg). ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline tokenizer = AutoTokenizer.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) model = AutoModelForSeq2SeqLM.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) text = "Replace by what you want." pipe = pipeline("text2text-generation", model=model, tokenizer=tokenizer, device=0) generated_text = pipe(text, truncation=True, max_length=64, no_repeat_ngram_size=7) ``` # ccdv/lsg-bart-base-4096-wcep This model is a fine-tuned version of [ccdv/lsg-bart-base-4096](https://huggingface.co/ccdv/lsg-bart-base-4096) on the [ccdv/WCEP-10 roberta](https://huggingface.co/datasets/ccdv/WCEP-10) dataset. \ It achieves the following results on the test set: | Length | Sparse Type | Block Size | Sparsity | Connexions | R1 | R2 | RL | RLsum | |:------ |:------------ |:---------- |:-------- | :--------- |:----- |:----- |:----- |:----- | | 4096 | Local | 256 | 0 | 768 | 46.02 | 24.23 | 37.38 | 38.72 | | 4096 | Local | 128 | 0 | 384 | 45.43 | 23.86 | 36.94 | 38.30 | | 4096 | Pooling | 128 | 4 | 644 | 45.36 | 23.61 | 36.75 | 38.06 | | 4096 | Stride | 128 | 4 | 644 | 45.87 | 24.31 | 37.41 | 38.70 | | 4096 | Block Stride | 128 | 4 | 644 | 45.78 | 24.16 | 37.20 | 38.48 | | 4096 | Norm | 128 | 4 | 644 | 45.34 | 23.39 | 36.47 | 37.78 | | 4096 | LSH | 128 | 4 | 644 | 45.15 | 23.53 | 36.74 | 38.02 | With smaller block size (lower ressources): | Length | Sparse Type | Block Size | Sparsity | Connexions | R1 | R2 | RL | RLsum | |:------ |:------------ |:---------- |:-------- | :--------- |:----- |:----- |:----- |:----- | | 4096 | Local | 64 | 0 | 192 | 44.48 | 22.98 | 36.20 | 37.52 | | 4096 | Local | 32 | 0 | 96 | 43.60 | 22.17 | 35.61 | 36.66 | | 4096 | Pooling | 32 | 4 | 160 | 43.91 | 22.41 | 35.80 | 36.92 | | 4096 | Stride | 32 | 4 | 160 | 44.62 | 23.11 | 36.32 | 37.53 | | 4096 | Block Stride | 32 | 4 | 160 | 44.47 | 23.02 | 36.28 | 37.46 | | 4096 | Norm | 32 | 4 | 160 | 44.45 | 23.03 | 36.10 | 37.33 | | 4096 | LSH | 32 | 4 | 160 | 43.87 | 22.50 | 35.75 | 36.93 | ## Model description The model relies on Local-Sparse-Global attention to handle long sequences: ![attn](attn.png) The model has about ~145 millions parameters (6 encoder layers - 6 decoder layers). \ The model is warm started from BART-base, converted to handle long sequences (encoder only) and fine tuned. ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 8e-05 - train_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 10.0 ### Generate hyperparameters The following hyperparameters were used during generation: - dataset_name: ccdv/WCEP-10 - dataset_config_name: roberta - eval_batch_size: 8 - eval_samples: 1022 - early_stopping: True - ignore_pad_token_for_loss: True - length_penalty: 2.0 - max_length: 64 - min_length: 0 - num_beams: 5 - no_repeat_ngram_size: None - seed: 123 ### Framework versions - Transformers 4.18.0 - Pytorch 1.10.1+cu102 - Datasets 2.1.0 - Tokenizers 0.11.6

null

Non_BioNLP

**Transformers >= 4.36.1**\ **This model relies on a custom modeling file, you need to add trust_remote_code=True**\ **See [\#13467](https://github.com/huggingface/transformers/pull/13467)** LSG ArXiv [paper](https://arxiv.org/abs/2210.15497). \ Github/conversion script is available at this [link](https://github.com/ccdv-ai/convert_checkpoint_to_lsg). ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline tokenizer = AutoTokenizer.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) model = AutoModelForSeq2SeqLM.from_pretrained("ccdv/lsg-bart-base-4096-wcep", trust_remote_code=True) text = "Replace by what you want." pipe = pipeline("text2text-generation", model=model, tokenizer=tokenizer, device=0) generated_text = pipe(text, truncation=True, max_length=64, no_repeat_ngram_size=7) ``` # ccdv/lsg-bart-base-4096-wcep This model is a fine-tuned version of [ccdv/lsg-bart-base-4096](https://huggingface.co/ccdv/lsg-bart-base-4096) on the [ccdv/WCEP-10 roberta](https://huggingface.co/datasets/ccdv/WCEP-10) dataset. \ It achieves the following results on the test set: | Length | Sparse Type | Block Size | Sparsity | Connexions | R1 | R2 | RL | RLsum | |:------ |:------------ |:---------- |:-------- | :--------- |:----- |:----- |:----- |:----- | | 4096 | Local | 256 | 0 | 768 | 46.02 | 24.23 | 37.38 | 38.72 | | 4096 | Local | 128 | 0 | 384 | 45.43 | 23.86 | 36.94 | 38.30 | | 4096 | Pooling | 128 | 4 | 644 | 45.36 | 23.61 | 36.75 | 38.06 | | 4096 | Stride | 128 | 4 | 644 | 45.87 | 24.31 | 37.41 | 38.70 | | 4096 | Block Stride | 128 | 4 | 644 | 45.78 | 24.16 | 37.20 | 38.48 | | 4096 | Norm | 128 | 4 | 644 | 45.34 | 23.39 | 36.47 | 37.78 | | 4096 | LSH | 128 | 4 | 644 | 45.15 | 23.53 | 36.74 | 38.02 | With smaller block size (lower ressources): | Length | Sparse Type | Block Size | Sparsity | Connexions | R1 | R2 | RL | RLsum | |:------ |:------------ |:---------- |:-------- | :--------- |:----- |:----- |:----- |:----- | | 4096 | Local | 64 | 0 | 192 | 44.48 | 22.98 | 36.20 | 37.52 | | 4096 | Local | 32 | 0 | 96 | 43.60 | 22.17 | 35.61 | 36.66 | | 4096 | Pooling | 32 | 4 | 160 | 43.91 | 22.41 | 35.80 | 36.92 | | 4096 | Stride | 32 | 4 | 160 | 44.62 | 23.11 | 36.32 | 37.53 | | 4096 | Block Stride | 32 | 4 | 160 | 44.47 | 23.02 | 36.28 | 37.46 | | 4096 | Norm | 32 | 4 | 160 | 44.45 | 23.03 | 36.10 | 37.33 | | 4096 | LSH | 32 | 4 | 160 | 43.87 | 22.50 | 35.75 | 36.93 | ## Model description The model relies on Local-Sparse-Global attention to handle long sequences: ![attn](attn.png) The model has about ~145 millions parameters (6 encoder layers - 6 decoder layers). \ The model is warm started from BART-base, converted to handle long sequences (encoder only) and fine tuned. ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 8e-05 - train_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 10.0 ### Generate hyperparameters The following hyperparameters were used during generation: - dataset_name: ccdv/WCEP-10 - dataset_config_name: roberta - eval_batch_size: 8 - eval_samples: 1022 - early_stopping: True - ignore_pad_token_for_loss: True - length_penalty: 2.0 - max_length: 64 - min_length: 0 - num_beams: 5 - no_repeat_ngram_size: None - seed: 123 ### Framework versions - Transformers 4.18.0 - Pytorch 1.10.1+cu102 - Datasets 2.1.0 - Tokenizers 0.11.6

{"datasets": ["ccdv/WCEP-10"], "language": ["en"], "metrics": ["rouge"], "tags": ["summarization"], "model-index": [{"name": "ccdv/lsg-bart-base-4096-wcep", "results": []}]}

task

[ "SUMMARIZATION" ]

44,630

Shankhdhar/ecommerce_query_classifier

Shankhdhar

text-classification

[ "sentence-transformers", "safetensors", "bert", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

2023-12-03T18:13:32Z

2023-12-03T18:25:32+00:00

50

0

--- license: apache-2.0 pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification --- #Shankhdhar/ecommerce_query_classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Shankhdhar/ecommerce_query_classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

null

Non_BioNLP

#Shankhdhar/ecommerce_query_classifier This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("Shankhdhar/ecommerce_query_classifier") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

{"license": "apache-2.0", "pipeline_tag": "text-classification", "tags": ["setfit", "sentence-transformers", "text-classification"]}

task

[ "TEXT_CLASSIFICATION" ]

44,631