{
  "profile": "gme2b",
  "data_dir": "/lus/lfs1aip2/scratch/u6ko/icywang.u6ko/mmrag_data",
  "pool": "built/pool_mix.jsonl",
  "corpus": "built/corpus_mix.jsonl",
  "image_tars": null,
  "image_dataset": "mix",
  "output_dir": "/lus/lfs1aip2/scratch/u6ko/icywang.u6ko/mmrag_data/runs/rl-j2e5-nogold-mix",
  "algo": "grpo",
  "temperature": 0.02,
  "learning_rate": 2e-05,
  "batch_size": 4,
  "num_candidates": 8,
  "max_steps": 500,
  "lora_r": 32,
  "lora_alpha": 64,
  "epsilon": 0.2,
  "vf_coef": 0.0,
  "value_head_lr": 0.01,
  "value_init": 0.3,
  "critic_warmup_steps": 20,
  "beta": 0.0,
  "contrastive_coef": 0.3,
  "contrastive_temperature": 0.03,
  "no_force_gold": true,
  "pool_sampling": false,
  "pool_sample_temperature": null,
  "reward_gate_std": 0.0,
  "reward": "judge",
  "slate_k": 5,
  "reward_server_url": null,
  "reader_model": "Qwen/Qwen2.5-VL-3B-Instruct",
  "reader_device": null,
  "reader_rollouts": 4,
  "reader_temperature": 0.7,
  "reader_batch_size": 16,
  "n_distractor_articles": 3000,
  "refresh_steps": 100,
  "encode_bs": 64,
  "max_len_doc": 512,
  "max_train_rows": 0,
  "query_instruction": null,
  "device": "cuda:0",
  "save_steps": 100,
  "log_every": 5,
  "seed": 0
}