| { | |
| "data_path": "/h/pritam/pritam_ssd004/OUTPUTS/Video-LLM/Phase2/GenData/video_chat2_mistral/tp_random-random_erase-stage2/diverse_10.json", | |
| "media_root": "", | |
| "num_frames": 16, | |
| "yes_count": 500, | |
| "no_count": 750, | |
| "num_workers": 6, | |
| "batch_size": 2, | |
| "gradient_accumulation_steps": 4, | |
| "trainable_modules": [], | |
| "subset_size": -1, | |
| "init_lora_zero": false, | |
| "model_name_or_path": "/h/pritam/pritam_ssd004/.cache/huggingface/hub/VideoChat2_stage3_Mistral_7B", | |
| "model": { | |
| "lora_r": 128, | |
| "lora_alpha": 256, | |
| "lora_dropout": 0.05, | |
| "lora_bias": "none" | |
| }, | |
| "loss_alpha": 0.01, | |
| "loss_beta": 0.9, | |
| "loss_fn": "phrase_dpo_tkld", | |
| "optimizer": { | |
| "opt": "adamW", | |
| "lr": 2e-05, | |
| "opt_betas": [ | |
| 0.9, | |
| 0.999 | |
| ], | |
| "weight_decay": 0.02, | |
| "max_grad_norm": -1, | |
| "different_lr": { | |
| "enable": false, | |
| "module_names": [], | |
| "lr": 0.001 | |
| } | |
| }, | |
| "scheduler": { | |
| "sched": "cosine", | |
| "epochs": 1, | |
| "min_lr_multi": 0.01, | |
| "warmup_epochs": 0.0 | |
| }, | |
| "fp16": true, | |
| "gradient_checkpointing": true, | |
| "wandb": { | |
| "enable": true, | |
| "entity": "pritamqu", | |
| "project": "ViSA-VideoChat2" | |
| }, | |
| "dist_url": "env://", | |
| "device": "cuda", | |
| "output_dir": "/h/pritam/pritam_ssd004/OUTPUTS/Video-LLM/Phase2/ViSA/videochat2_mistral_7b/202503070207031433_phrase_dpo_tkld_2e-5_0.01_0.9_500_750_1_diverse_10_tp_random-random_erase_lora_128_256_cosine_lr_gradfix_dynamic", | |
| "resume": false, | |
| "debug": false, | |
| "log_freq": 1, | |
| "seed": 42, | |
| "save_latest": true, | |
| "auto_resume": true, | |
| "pretrained_path": "", | |
| "rank": 0, | |
| "world_size": 4, | |
| "gpu": 0, | |
| "distributed": true, | |
| "dist_backend": "nccl" | |
| } |