AAAzsf/HDCEval_PQR

1exp_name: mae2subdomain: pqr3datasets:4- medical-eval5- medical-meadow-wikidoc6- med-text7- med-dialogue8dataset_directory: Model_Training/Dataset/train9evaluation_model_priority:10- gpt411- gpt3512n_epochs: 113n_examples: null14n_eval_examples: 1015batch_size: 12816gradient_accumulation_steps: 6417eval_batch_size: 218max_length: 352019max_prompt_length: 292420reward_token: Unified21frozen_layers:22  enable: true23  language_model_layers:24  - '0'25  - '1'26  - '2'27  - '3'28  - '4'29  - '5'30  - '6'31  - '7'32  - '8'33  - '9'34  - '10'35  - '11'36  - '12'37  - '13'38  - '14'39  - '15'40  - '16'41  - '17'42  - '18'43  - '19'44  - '20'45  - '21'46  - '22'47  - '23'48  vision_tower_layers: -149debug: false50fsdp_port: 5706751wandb:52  enabled: true53  entity: null54  project: direct-preference-optimization55local_dirs:56- /scr-ssd57- /scr58- .cache59sample_during_eval: false60n_eval_model_samples: 261do_first_eval: true62local_run_dir: .cache/lhe_exc/mae_2024-04-18_21-20-49_10858963lr: 2.0e-0664max_grad_norm: 10.065trainer: FSDPTrainer66optimizer: RMSprop67warmup_steps: 1568activation_checkpointing: false69eval_every: 1497670minimum_log_interval_secs: 1.071seed: 072model:73  name_or_path: ./LM/medllama2/74  tokenizer_name_or_path: null75  archive: ./LM/mae/mae_pqr/sft_only/Unified/policy.pt76  block_name: LlamaDecoderLayer77  policy_dtype: bfloat1678  fsdp_policy_mp: bfloat1679  reference_dtype: bfloat1680loss:81  name: dpo82  beta: 0.183  reference_free: false84