{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8888888888888888,
  "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": 477.0,
      "epoch": 0.08888888888888889,
      "grad_norm": 34.471927642822266,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": -0.0,
      "reward": 1.174295978154987,
      "reward_std": 1.7542744056117954,
      "rewards/concensus_correctness_reward_func": 0.8310000002384186,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.17798347567440942,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.015625,
      "rewards/xmlcount_reward_func": 0.08718749997206032,
      "step": 2
    },
    {
      "completion_length": 277.25,
      "epoch": 0.17777777777777778,
      "grad_norm": 26.461362838745117,
      "kl": 0.00166156359409797,
      "learning_rate": 4.864543104251586e-07,
      "loss": 0.0,
      "reward": 0.6097803856246173,
      "reward_std": 0.5241530751809478,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.125,
      "rewards/question_recreation_reward_func": 0.2592491409741342,
      "rewards/soft_format_reward_func": 0.015625,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.20990624744445086,
      "step": 4
    },
    {
      "completion_length": 386.34375,
      "epoch": 0.26666666666666666,
      "grad_norm": 11.892231941223145,
      "kl": 0.0016316118271788582,
      "learning_rate": 4.472851273490984e-07,
      "loss": 0.0,
      "reward": 0.9368757626507431,
      "reward_std": 1.3130901722470298,
      "rewards/concensus_correctness_reward_func": 0.1224374994635582,
      "rewards/consensus_reward_func": 0.125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.3125,
      "rewards/question_recreation_reward_func": 0.2546570213744417,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.12228124856483191,
      "step": 6
    },
    {
      "completion_length": 324.9375,
      "epoch": 0.35555555555555557,
      "grad_norm": 22.727386474609375,
      "kl": 0.0020563425605359953,
      "learning_rate": 3.867370395306068e-07,
      "loss": 0.0,
      "reward": 2.560941856354475,
      "reward_std": 1.6560740754939616,
      "rewards/concensus_correctness_reward_func": 1.875,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.25,
      "rewards/question_recreation_reward_func": 0.258441973477602,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.015625,
      "rewards/xmlcount_reward_func": 0.16187499649822712,
      "step": 8
    },
    {
      "completion_length": 466.5625,
      "epoch": 0.4444444444444444,
      "grad_norm": 4.7971391677856445,
      "kl": 0.0010438499575684546,
      "learning_rate": 3.1137137178519977e-07,
      "loss": 0.0,
      "reward": 0.5703212053049356,
      "reward_std": 0.8992374676745385,
      "rewards/concensus_correctness_reward_func": 0.12262500077486038,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.125,
      "rewards/question_recreation_reward_func": 0.19388369610533118,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.06631250027567148,
      "step": 10
    },
    {
      "completion_length": 476.96875,
      "epoch": 0.5333333333333333,
      "grad_norm": 10.955451011657715,
      "kl": 0.0013942485475126887,
      "learning_rate": 2.2935516363191693e-07,
      "loss": 0.0,
      "reward": 0.5070485147298314,
      "reward_std": 0.39075206633424386,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.1875,
      "rewards/question_recreation_reward_func": 0.198392266232986,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.1211562491953373,
      "step": 12
    },
    {
      "completion_length": 502.15625,
      "epoch": 0.6222222222222222,
      "grad_norm": 5.031079292297363,
      "kl": 0.0016764367901487276,
      "learning_rate": 1.4957614383675767e-07,
      "loss": 0.0,
      "reward": 0.3395441100001335,
      "reward_std": 0.43814076797571033,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.3125,
      "rewards/question_recreation_reward_func": 0.1331691121449694,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.10612499457783997,
      "step": 14
    },
    {
      "completion_length": 339.5,
      "epoch": 0.7111111111111111,
      "grad_norm": 10.80864143371582,
      "kl": 0.0050469062916818075,
      "learning_rate": 8.067960709356478e-08,
      "loss": 0.0,
      "reward": 0.33586260332958773,
      "reward_std": 0.4518825721988833,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.125,
      "rewards/question_recreation_reward_func": 0.1327063522185199,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.015625,
      "rewards/xmlcount_reward_func": 0.06253124959766865,
      "step": 16
    },
    {
      "completion_length": 362.3125,
      "epoch": 0.8,
      "grad_norm": 15.09556770324707,
      "kl": 0.0029296423817868344,
      "learning_rate": 3.013156219837776e-08,
      "loss": 0.0,
      "reward": 1.0260260263457894,
      "reward_std": 1.464539691573009,
      "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": 0.24865100963506848,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.02737499843351543,
      "step": 18
    },
    {
      "completion_length": 406.875,
      "epoch": 0.8888888888888888,
      "grad_norm": 73.40243530273438,
      "kl": 0.006029042899172055,
      "learning_rate": 3.4096741493194193e-09,
      "loss": 0.0,
      "reward": 1.030112552922219,
      "reward_std": 1.194272859574994,
      "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": 0.16539380967151374,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.11471875058487058,
      "step": 20
    },
    {
      "epoch": 0.8888888888888888,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 2.187862992286682e-06,
      "train_runtime": 301.973,
      "train_samples_per_second": 1.06,
      "train_steps_per_second": 0.066
    }
  ],
  "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
}