{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 371.625,
      "epoch": 0.1,
      "grad_norm": 11.70975112915039,
      "kl": 0.0,
      "learning_rate": 4.965903258506806e-07,
      "loss": -0.0,
      "reward": 0.6997022139839828,
      "reward_std": 0.7695410805754364,
      "rewards/concensus_correctness_reward_func": 0.07256250083446503,
      "rewards/consensus_reward_func": 0.1875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.28182722203200683,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.15781249664723873,
      "step": 2
    },
    {
      "completion_length": 435.84375,
      "epoch": 0.2,
      "grad_norm": 7.437254428863525,
      "kl": 0.019949616707890527,
      "learning_rate": 4.698684378016222e-07,
      "loss": 0.0,
      "reward": 0.5647699106484652,
      "reward_std": 0.5821361623238772,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.125,
      "rewards/question_recreation_reward_func": 0.3737699074554257,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.003500005230307579,
      "step": 4
    },
    {
      "completion_length": 480.6875,
      "epoch": 0.3,
      "grad_norm": 7.619124889373779,
      "kl": 0.025194359630404506,
      "learning_rate": 4.193203929064353e-07,
      "loss": 0.0,
      "reward": 0.47773217875510454,
      "reward_std": 0.6364234593929723,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.32951344281900674,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.14821874536573887,
      "step": 6
    },
    {
      "completion_length": 415.96875,
      "epoch": 0.4,
      "grad_norm": 115.83442687988281,
      "kl": 0.13178696436079917,
      "learning_rate": 3.5042385616324236e-07,
      "loss": 0.0001,
      "reward": 0.6815493167378008,
      "reward_std": 0.7156448389869183,
      "rewards/concensus_correctness_reward_func": 0.05862500052899122,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.47239306825213134,
      "rewards/soft_format_reward_func": 0.015625,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.13490625086706132,
      "step": 8
    },
    {
      "completion_length": 411.5625,
      "epoch": 0.5,
      "grad_norm": 6.062715530395508,
      "kl": 0.01802918140310794,
      "learning_rate": 2.706448363680831e-07,
      "loss": 0.0,
      "reward": 0.8854871927178465,
      "reward_std": 0.559852755832253,
      "rewards/concensus_correctness_reward_func": 0.032500000670552254,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.3651746995165013,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.36281250417232513,
      "step": 10
    },
    {
      "completion_length": 352.65625,
      "epoch": 0.6,
      "grad_norm": 7.630370616912842,
      "kl": 0.004244693569489755,
      "learning_rate": 1.886286282148002e-07,
      "loss": 0.0,
      "reward": 0.6533340434543788,
      "reward_std": 0.5752735425485298,
      "rewards/concensus_correctness_reward_func": 0.07443749904632568,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.30983403185382485,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.20656249369494617,
      "step": 12
    },
    {
      "completion_length": 383.1875,
      "epoch": 0.7,
      "grad_norm": 4.60897159576416,
      "kl": 0.0038806117881904356,
      "learning_rate": 1.1326296046939333e-07,
      "loss": 0.0,
      "reward": 1.2285354160703719,
      "reward_std": 1.487802633550018,
      "rewards/concensus_correctness_reward_func": 0.625,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.125,
      "rewards/question_recreation_reward_func": 0.31894167221616954,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.1595937586389482,
      "step": 14
    },
    {
      "completion_length": 320.3125,
      "epoch": 0.8,
      "grad_norm": 6.664931297302246,
      "kl": 0.00507382119394606,
      "learning_rate": 5.271487265090163e-08,
      "loss": 0.0,
      "reward": 0.9667287697084248,
      "reward_std": 0.4696869046892971,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.25,
      "rewards/question_recreation_reward_func": 0.40541626675985754,
      "rewards/soft_format_reward_func": 0.015625,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.2956874957308173,
      "step": 16
    },
    {
      "completion_length": 357.90625,
      "epoch": 0.9,
      "grad_norm": 6.452459812164307,
      "kl": 0.01000264076719759,
      "learning_rate": 1.3545689574841341e-08,
      "loss": 0.0,
      "reward": 0.4444698728621006,
      "reward_std": 0.9062960698502138,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.3955323635600507,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.013562485575675964,
      "step": 18
    },
    {
      "completion_length": 341.0,
      "epoch": 1.0,
      "grad_norm": 8.592371940612793,
      "kl": 0.005512678173545282,
      "learning_rate": 0.0,
      "loss": 0.0,
      "reward": 1.0551672836299986,
      "reward_std": 1.034868648624979,
      "rewards/concensus_correctness_reward_func": 0.02656250074505806,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.25,
      "rewards/question_recreation_reward_func": 0.41138604609295726,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.3047187514603138,
      "step": 20
    },
    {
      "epoch": 1.0,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 2.2401101887226105e-05,
      "train_runtime": 228.7273,
      "train_samples_per_second": 1.399,
      "train_steps_per_second": 0.087
    }
  ],
  "logging_steps": 2,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}