{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "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": 344.5,
      "epoch": 0.20253164556962025,
      "grad_norm": 9.9772310256958,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": -0.0,
      "reward": 0.8219375633634627,
      "reward_std": 0.7066538096405566,
      "rewards/concensus_correctness_reward_func": 0.05999999865889549,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.2743438226170838,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.015625,
      "rewards/xmlcount_reward_func": 0.4094687553588301,
      "step": 2
    },
    {
      "completion_length": 377.8125,
      "epoch": 0.4050632911392405,
      "grad_norm": 8.212695121765137,
      "kl": 0.0010669995590433246,
      "learning_rate": 4.864543104251586e-07,
      "loss": 0.0,
      "reward": 0.9275585329160094,
      "reward_std": 0.9481612900854088,
      "rewards/concensus_correctness_reward_func": 0.06237500160932541,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.3217460191808641,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.4184375060722232,
      "step": 4
    },
    {
      "completion_length": 307.4375,
      "epoch": 0.6075949367088608,
      "grad_norm": 9.819817543029785,
      "kl": 0.001468627218855545,
      "learning_rate": 4.472851273490984e-07,
      "loss": 0.0,
      "reward": 0.9682412928668782,
      "reward_std": 0.59242624102626,
      "rewards/concensus_correctness_reward_func": 0.17887500301003456,
      "rewards/consensus_reward_func": 0.1875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.29499130306066945,
      "rewards/soft_format_reward_func": 0.015625,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.2912500018719584,
      "step": 6
    },
    {
      "completion_length": 366.21875,
      "epoch": 0.810126582278481,
      "grad_norm": 6.490735054016113,
      "kl": 0.0012780786164512392,
      "learning_rate": 3.867370395306068e-07,
      "loss": 0.0,
      "reward": 0.5933836810290813,
      "reward_std": 1.0359376267297193,
      "rewards/concensus_correctness_reward_func": 0.12250000238418579,
      "rewards/consensus_reward_func": 0.125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.24210240540560335,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.10378125216811895,
      "step": 8
    },
    {
      "completion_length": 375.8333333333333,
      "epoch": 1.0,
      "grad_norm": 6.2283196449279785,
      "kl": 0.001403550726051132,
      "learning_rate": 3.1137137178519977e-07,
      "loss": 0.0,
      "reward": 0.9483249210442106,
      "reward_std": 0.7847038037454088,
      "rewards/concensus_correctness_reward_func": 0.0665333350499471,
      "rewards/consensus_reward_func": 0.06666666666666667,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.06666666666666667,
      "rewards/question_recreation_reward_func": 0.35792491491883993,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.3905333320299784,
      "step": 10
    },
    {
      "completion_length": 509.8125,
      "epoch": 1.2025316455696202,
      "grad_norm": 6.293169975280762,
      "kl": 0.0016378157160943374,
      "learning_rate": 2.2935516363191693e-07,
      "loss": 0.0,
      "reward": 0.40420354809612036,
      "reward_std": 1.324048607615623,
      "rewards/concensus_correctness_reward_func": 0.12250000238418579,
      "rewards/consensus_reward_func": 0.125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.23114103089028504,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.07443749927915633,
      "step": 12
    },
    {
      "completion_length": 362.4375,
      "epoch": 1.4050632911392404,
      "grad_norm": 9.059679985046387,
      "kl": 0.002234285893791821,
      "learning_rate": 1.4957614383675767e-07,
      "loss": 0.0,
      "reward": 0.6384785049594939,
      "reward_std": 0.7130910204723477,
      "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": 0.3409159923903644,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.23506250511854887,
      "step": 14
    },
    {
      "completion_length": 257.625,
      "epoch": 1.6075949367088609,
      "grad_norm": 17.0372314453125,
      "kl": 0.0024873377551557496,
      "learning_rate": 8.067960709356478e-08,
      "loss": 0.0,
      "reward": 0.7720209148246795,
      "reward_std": 0.5560847400847706,
      "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": 0.3203334001591429,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.38918749894946814,
      "step": 16
    },
    {
      "completion_length": 436.65625,
      "epoch": 1.810126582278481,
      "grad_norm": 6.472311973571777,
      "kl": 0.0021311557193257613,
      "learning_rate": 3.013156219837776e-08,
      "loss": 0.0,
      "reward": 0.6501477827550843,
      "reward_std": 0.8022108940640464,
      "rewards/concensus_correctness_reward_func": 0.05412499979138374,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.2959602870978415,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.17506249761208892,
      "step": 18
    },
    {
      "completion_length": 405.1,
      "epoch": 2.0,
      "grad_norm": 13.560686111450195,
      "kl": 0.0018062964198179543,
      "learning_rate": 3.4096741493194193e-09,
      "loss": 0.0,
      "reward": 0.7393059747914473,
      "reward_std": 0.6481222245842219,
      "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": 0.428839307030042,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.31046666838228704,
      "step": 20
    },
    {
      "epoch": 2.0,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 1.4645672763435869e-06,
      "train_runtime": 1759.1185,
      "train_samples_per_second": 0.182,
      "train_steps_per_second": 0.011
    }
  ],
  "logging_steps": 2,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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
}