{
  "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": 137.28125,
      "epoch": 0.5714285714285714,
      "grad_norm": 4.629494667053223,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 4.0658501498401165,
      "reward_std": 0.9820615029020701,
      "rewards/concensus_correctness_reward_func": 1.2021875157952309,
      "rewards/consensus_reward_func": 1.25,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.8004126728046685,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7507500052452087,
      "step": 2
    },
    {
      "completion_length": 157.58333333333334,
      "epoch": 1.0,
      "grad_norm": 5.000041961669922,
      "kl": 0.013368819782044739,
      "learning_rate": 4.864543104251586e-07,
      "loss": 0.0,
      "reward": 5.314966678619385,
      "reward_std": 0.8371213575204214,
      "rewards/concensus_correctness_reward_func": 2.0760000149408975,
      "rewards/consensus_reward_func": 1.8333333333333333,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.16666666666666666,
      "rewards/question_recreation_reward_func": 0.4317583265559127,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.807208334406217,
      "step": 4
    },
    {
      "completion_length": 114.125,
      "epoch": 1.5714285714285714,
      "grad_norm": 4.73675537109375,
      "kl": 0.01866700834580115,
      "learning_rate": 4.472851273490984e-07,
      "loss": 0.0,
      "reward": 4.9177690201904625,
      "reward_std": 0.5806675790809095,
      "rewards/concensus_correctness_reward_func": 1.7927500158548355,
      "rewards/consensus_reward_func": 1.75,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.44886276265606284,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.8636562526226044,
      "step": 6
    },
    {
      "completion_length": 132.625,
      "epoch": 2.0,
      "grad_norm": 2.4207491874694824,
      "kl": 0.01655085780657828,
      "learning_rate": 3.867370395306068e-07,
      "loss": 0.0,
      "reward": 5.619066794713338,
      "reward_std": 0.5693128562609976,
      "rewards/concensus_correctness_reward_func": 2.00966669122378,
      "rewards/consensus_reward_func": 1.9166666666666667,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.08333333333333333,
      "rewards/question_recreation_reward_func": 0.8152334044377009,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7941666742165884,
      "step": 8
    },
    {
      "completion_length": 138.1875,
      "epoch": 2.571428571428571,
      "grad_norm": 4.383825778961182,
      "kl": 0.017200405127368867,
      "learning_rate": 3.1137137178519977e-07,
      "loss": 0.0,
      "reward": 5.2846749955788255,
      "reward_std": 0.34504761872813106,
      "rewards/concensus_correctness_reward_func": 1.852875016629696,
      "rewards/consensus_reward_func": 1.8125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.125,
      "rewards/question_recreation_reward_func": 0.722112582763657,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7721874918788671,
      "step": 10
    },
    {
      "completion_length": 122.66666666666667,
      "epoch": 3.0,
      "grad_norm": 1.8496273756027222,
      "kl": 0.015685971166628104,
      "learning_rate": 2.2935516363191693e-07,
      "loss": 0.0,
      "reward": 5.725431720415751,
      "reward_std": 0.4189778131355221,
      "rewards/concensus_correctness_reward_func": 2.00966669122378,
      "rewards/consensus_reward_func": 1.9166666666666667,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.16666666666666666,
      "rewards/question_recreation_reward_func": 0.8341400871674219,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7982916633288065,
      "step": 12
    },
    {
      "completion_length": 144.09375,
      "epoch": 3.571428571428571,
      "grad_norm": 5.370832443237305,
      "kl": 0.008550927566830069,
      "learning_rate": 1.4957614383675767e-07,
      "loss": 0.0,
      "reward": 5.755253612995148,
      "reward_std": 0.11208663968136534,
      "rewards/concensus_correctness_reward_func": 2.0373750180006027,
      "rewards/consensus_reward_func": 1.875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.25,
      "rewards/question_recreation_reward_func": 0.8512537020724267,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7416250053793192,
      "step": 14
    },
    {
      "completion_length": 129.625,
      "epoch": 4.0,
      "grad_norm": 2.3071887493133545,
      "kl": 0.011528774998926869,
      "learning_rate": 8.067960709356478e-08,
      "loss": 0.0,
      "reward": 5.503378629684448,
      "reward_std": 0.1692266787867993,
      "rewards/concensus_correctness_reward_func": 1.9240000247955322,
      "rewards/consensus_reward_func": 2.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.8247120107213656,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7546666612227758,
      "step": 16
    },
    {
      "completion_length": 158.8125,
      "epoch": 4.571428571428571,
      "grad_norm": 4.723745822906494,
      "kl": 0.011602169910474913,
      "learning_rate": 3.013156219837776e-08,
      "loss": 0.0,
      "reward": 5.4294650866650045,
      "reward_std": 0.24022297459305264,
      "rewards/concensus_correctness_reward_func": 1.917125016450882,
      "rewards/consensus_reward_func": 1.75,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.1875,
      "rewards/question_recreation_reward_func": 0.7994339219294488,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.775406252592802,
      "step": 18
    },
    {
      "completion_length": 135.04166666666666,
      "epoch": 5.0,
      "grad_norm": 1.8162436485290527,
      "kl": 0.03148796226014383,
      "learning_rate": 3.4096741493194193e-09,
      "loss": 0.0,
      "reward": 5.291161974271138,
      "reward_std": 0.6445521160882587,
      "rewards/concensus_correctness_reward_func": 1.7636666893959045,
      "rewards/consensus_reward_func": 1.8333333333333333,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.9020370791355768,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.7921250065167745,
      "step": 20
    },
    {
      "epoch": 5.0,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 1.201600916829193e-05,
      "train_runtime": 14036.781,
      "train_samples_per_second": 0.023,
      "train_steps_per_second": 0.001
    }
  ],
  "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
}