{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "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": 326.59375,
      "epoch": 0.5714285714285714,
      "grad_norm": 4.934439182281494,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 0.9843985685729422,
      "reward_std": 0.7370969671756029,
      "rewards/concensus_correctness_reward_func": -0.02437499910593033,
      "rewards/consensus_reward_func": 0.25,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.39177357318112627,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.3670000056736171,
      "step": 2
    },
    {
      "completion_length": 371.0,
      "epoch": 1.0,
      "grad_norm": 3.6815781593322754,
      "kl": 0.0011949266287653397,
      "learning_rate": 4.864543104251586e-07,
      "loss": 0.0,
      "reward": 1.5001778639853,
      "reward_std": 0.7827915487190088,
      "rewards/concensus_correctness_reward_func": 0.03458333263794581,
      "rewards/consensus_reward_func": 0.25,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.25,
      "rewards/question_recreation_reward_func": 0.4009695214529832,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.5646250148614248,
      "step": 4
    },
    {
      "completion_length": 359.96875,
      "epoch": 1.5714285714285714,
      "grad_norm": 4.556028842926025,
      "kl": 0.0010278360459778924,
      "learning_rate": 4.472851273490984e-07,
      "loss": 0.0,
      "reward": 1.1807272167643532,
      "reward_std": 0.653216159567819,
      "rewards/concensus_correctness_reward_func": 0.06068750098347664,
      "rewards/consensus_reward_func": 0.125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.47382097283843905,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.45871875062584877,
      "step": 6
    },
    {
      "completion_length": 439.2916666666667,
      "epoch": 2.0,
      "grad_norm": 2.2755231857299805,
      "kl": 0.0013096223750229303,
      "learning_rate": 3.867370395306068e-07,
      "loss": 0.0,
      "reward": 0.7530886133511862,
      "reward_std": 0.9865690590813756,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.08333333333333333,
      "rewards/question_recreation_reward_func": 0.37442193475241464,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.295333335796992,
      "step": 8
    },
    {
      "completion_length": 370.40625,
      "epoch": 2.571428571428571,
      "grad_norm": 6.62834358215332,
      "kl": 0.0012265997065696865,
      "learning_rate": 3.1137137178519977e-07,
      "loss": 0.0,
      "reward": 0.8175993217155337,
      "reward_std": 1.0472558625042439,
      "rewards/concensus_correctness_reward_func": 0.05756250023841858,
      "rewards/consensus_reward_func": 0.125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.4674430672894232,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.16759375017136335,
      "step": 10
    },
    {
      "completion_length": 356.4583333333333,
      "epoch": 3.0,
      "grad_norm": 3.410081148147583,
      "kl": 0.001667177789689352,
      "learning_rate": 2.2935516363191693e-07,
      "loss": 0.0,
      "reward": 1.0313266267379124,
      "reward_std": 0.6612533178801338,
      "rewards/concensus_correctness_reward_func": 0.0790833334128062,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.572909953383108,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.37933334335684776,
      "step": 12
    },
    {
      "completion_length": 401.03125,
      "epoch": 3.571428571428571,
      "grad_norm": 6.091869354248047,
      "kl": 0.0018705126858549193,
      "learning_rate": 1.4957614383675767e-07,
      "loss": 0.0,
      "reward": 0.7606599755818024,
      "reward_std": 0.9554330813698471,
      "rewards/concensus_correctness_reward_func": 0.06806249916553497,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.5348162406007759,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.09528125822544098,
      "step": 14
    },
    {
      "completion_length": 407.9166666666667,
      "epoch": 4.0,
      "grad_norm": 2.649594306945801,
      "kl": 0.001604035545218115,
      "learning_rate": 8.067960709356478e-08,
      "loss": 0.0,
      "reward": 0.6916571507851282,
      "reward_std": 1.01990426153255,
      "rewards/concensus_correctness_reward_func": 0.024666666984558105,
      "rewards/consensus_reward_func": 0.08333333333333333,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.4062821501865983,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.17737499624490738,
      "step": 16
    },
    {
      "completion_length": 412.125,
      "epoch": 4.571428571428571,
      "grad_norm": 119.07859802246094,
      "kl": 0.0031047585580381565,
      "learning_rate": 3.013156219837776e-08,
      "loss": 0.0,
      "reward": 0.8601471782312728,
      "reward_std": 0.6142369294539094,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.464647185173817,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.2705000061541796,
      "step": 18
    },
    {
      "completion_length": 346.3333333333333,
      "epoch": 5.0,
      "grad_norm": 4.018592834472656,
      "kl": 0.0016965343481084953,
      "learning_rate": 3.4096741493194193e-09,
      "loss": 0.0,
      "reward": 1.0170780109862487,
      "reward_std": 0.9925470755745968,
      "rewards/concensus_correctness_reward_func": -0.0019999990860621133,
      "rewards/consensus_reward_func": 0.25,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.08333333333333333,
      "rewards/question_recreation_reward_func": 0.32816133989642066,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.35758334149916965,
      "step": 20
    },
    {
      "epoch": 5.0,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 1.2790784239768983e-06,
      "train_runtime": 291.7059,
      "train_samples_per_second": 1.097,
      "train_steps_per_second": 0.069
    }
  ],
  "logging_steps": 2,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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
}