%%bash CUDA_VISIBLE_DEVICES=0
MAX_PIXELS=1003520
swift rlhf
--rlhf_type grpo
--model 'Mihaiii/Ovis2-4B'
--model_type ovis2
--external_plugins './plugin.py'
--reward_funcs external_db_accuracy
--train_type lora
--lora_rank 8
--lora_alpha 32
--target_modules all-linear
--torch_dtype bfloat16
--dataset 'Mihaiii/OCRBenchv2-other-cols-2'
--use_hf 1
--max_completion_length 1024
--num_train_epochs 1
--per_device_train_batch_size 4
--per_device_eval_batch_size 4
--learning_rate 1e-4
--gradient_accumulation_steps 1
--eval_steps 50
--save_steps 250
--save_total_limit 2
--logging_steps 5
--max_length 2048
--output_dir output
--warmup_ratio 0.05
--dataloader_num_workers 4
--dataset_num_proc 4
--num_generations 4
--temperature 0.9
--split_dataset_ratio 0

Downloads last month
10
Safetensors
Model size
4.3B params
Tensor type
BF16
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The HF Inference API does not support model that require custom code execution.

Space using Mihaiii/Ovis2-4B-RL-VQA-1 1