gavinqiangli/bge-base-financial-matryoshka

BGE base Financial Matryoshka

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 on the json dataset. 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
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json
• Language: en
• License: apache-2.0

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

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:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("gavinqiangli/bge-base-financial-matryoshka")
# Run inference
sentences = [
    'In 2023, total assets associated with derivatives designated as hedging instruments amounted to $1,527 million, while total liabilities amounted to $5,962 million.',
    'What was the total value of assets and liabilities associated with derivatives designated as hedging instruments in 2023?',
    'What was the balance of deferred net loss on derivatives included in accumulated other comprehensive income as of December 31, 2023?',
]
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_768
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.7314
cosine_accuracy@3	0.8414
cosine_accuracy@5	0.88
cosine_accuracy@10	0.9171
cosine_precision@1	0.7314
cosine_precision@3	0.2805
cosine_precision@5	0.176
cosine_precision@10	0.0917
cosine_recall@1	0.7314
cosine_recall@3	0.8414
cosine_recall@5	0.88
cosine_recall@10	0.9171
cosine_ndcg@10	0.8243
cosine_mrr@10	0.7946
cosine_map@100	0.7974

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.73
cosine_accuracy@3	0.8443
cosine_accuracy@5	0.8757
cosine_accuracy@10	0.9114
cosine_precision@1	0.73
cosine_precision@3	0.2814
cosine_precision@5	0.1751
cosine_precision@10	0.0911
cosine_recall@1	0.73
cosine_recall@3	0.8443
cosine_recall@5	0.8757
cosine_recall@10	0.9114
cosine_ndcg@10	0.8208
cosine_mrr@10	0.7918
cosine_map@100	0.7951

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.7243
cosine_accuracy@3	0.8286
cosine_accuracy@5	0.8671
cosine_accuracy@10	0.9057
cosine_precision@1	0.7243
cosine_precision@3	0.2762
cosine_precision@5	0.1734
cosine_precision@10	0.0906
cosine_recall@1	0.7243
cosine_recall@3	0.8286
cosine_recall@5	0.8671
cosine_recall@10	0.9057
cosine_ndcg@10	0.8145
cosine_mrr@10	0.7854
cosine_map@100	0.7887

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.7043
cosine_accuracy@3	0.8129
cosine_accuracy@5	0.8557
cosine_accuracy@10	0.9057
cosine_precision@1	0.7043
cosine_precision@3	0.271
cosine_precision@5	0.1711
cosine_precision@10	0.0906
cosine_recall@1	0.7043
cosine_recall@3	0.8129
cosine_recall@5	0.8557
cosine_recall@10	0.9057
cosine_ndcg@10	0.8019
cosine_mrr@10	0.769
cosine_map@100	0.7722

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6657
cosine_accuracy@3	0.78
cosine_accuracy@5	0.82
cosine_accuracy@10	0.8686
cosine_precision@1	0.6657
cosine_precision@3	0.26
cosine_precision@5	0.164
cosine_precision@10	0.0869
cosine_recall@1	0.6657
cosine_recall@3	0.78
cosine_recall@5	0.82
cosine_recall@10	0.8686
cosine_ndcg@10	0.7668
cosine_mrr@10	0.7344
cosine_map@100	0.7398

Training Details

Training Dataset

json

• Dataset: json
• Size: 6,300 training samples
• Columns: positive and anchor
• Approximate statistics based on the first 1000 samples:

  	positive	anchor
  type	string	string
  details	min: 10 tokens mean: 46.04 tokens max: 289 tokens	min: 2 tokens mean: 20.38 tokens max: 42 tokens

• Samples:

  positive	anchor
  The Nominating and Corporate Governance Committee of our Board of Directors is responsible for reviewing and discussing with management our practices related to ESG.	What is the role of the Nominating and Corporate Governance Committee at NVIDIA?
  Deferred tax assets and deferred tax liabilities included in the Consolidated Balance Sheets as follows: As of October 31, 2023: Deferred tax assets were $3,155 million and Deferred tax liabilities were $44 million. As of October 31, 2022: Deferred tax assets were $2,167 million and Deferred tax liabilities were $121 million. The total net deferred tax assets were $3,111 million in 2023 and $2,046 million in 2022.	What was the change in HP's net deferred tax assets from 2022 to 2023?
  Sales and marketing expense increased $247 million, or 16%, in 2023, compared to 2022, primarily due to a $177 million increase in marketing activities associated with our marketing campaigns and launches and our search engine marketing and advertising spend.	What was the major reason for the increase in Sales and Marketing expenses in 2023?

• Loss: MatryoshkaLoss with these parameters:

  {
      "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: 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
• fp16: True
• load_best_model_at_end: True
• optim: adamw_torch_fused
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• 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: 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_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

Training Logs

Epoch	Step	Training Loss	dim_768_cosine_map@100	dim_512_cosine_map@100	dim_256_cosine_map@100	dim_128_cosine_map@100	dim_64_cosine_map@100
0.8122	10	1.5963	-	-	-	-	-
0.9746	12	-	0.7791	0.7824	0.7662	0.7483	0.7086
1.6244	20	0.6846	-	-	-	-	-
1.9492	24	-	0.7924	0.7903	0.7859	0.7664	0.7327
2.4365	30	0.4956	-	-	-	-	-
2.9239	36	-	0.7962	0.7939	0.7886	0.7716	0.7378
3.2487	40	0.3998	-	-	-	-	-
3.8985	48	-	0.7974	0.7951	0.7887	0.7722	0.7398

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.2.1
• Transformers: 4.41.2
• PyTorch: 2.1.2+cu121
• Accelerate: 0.34.2
• Datasets: 2.19.1
• Tokenizers: 0.19.1

Citation

BibTeX

Sentence Transformers

@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

@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

@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}
}