| ### model | |
| model_name_or_path: Qwen/Qwen2.5-VL-3B-Instruct | |
| image_max_pixels: 12845056 | |
| video_max_pixels: 16384 | |
| trust_remote_code: true | |
| ### method | |
| stage: sft | |
| do_train: true | |
| finetuning_type: full | |
| freeze_vision_tower: false # choices: [true, false] | |
| freeze_multi_modal_projector: false # choices: [true, false] | |
| train_mm_proj_only: false # choices: [true, false] | |
| deepspeed: examples/deepspeed/ds_z2_config.json # choices: [ds_z0_config.json, ds_z2_config.json, ds_z3_config.json] | |
| ### dataset | |
| dataset: mllm_rec_json | |
| template: qwen2_vl | |
| cutoff_len: 131072 | |
| max_samples: 10000000 | |
| overwrite_cache: true | |
| preprocessing_num_workers: 16 | |
| ### output | |
| output_dir: saves/qwen2_5_vl-3b/full/sft | |
| logging_steps: 10 | |
| save_steps: 100 | |
| plot_loss: true | |
| overwrite_output_dir: true | |
| ### train | |
| per_device_train_batch_size: 1 | |
| gradient_accumulation_steps: 2 | |
| learning_rate: 1.0e-6 | |
| num_train_epochs: 2.0 | |
| lr_scheduler_type: cosine | |
| warmup_ratio: 0.1 | |
| bf16: true | |
| ddp_timeout: 180000000 |