{
  "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": 4.0,
      "epoch": 0.1,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 3.9487337917089462,
      "reward_std": 0.0,
      "rewards/concensus_correctness_reward_func": 1.9237500205636024,
      "rewards/consensus_reward_func": 2.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.024983764393255115,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 2
    },
    {
      "completion_length": 4.0,
      "epoch": 0.2,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 4.864543104251586e-07,
      "loss": 0.0,
      "reward": 3.833080992102623,
      "reward_std": 0.0,
      "rewards/concensus_correctness_reward_func": 2.047750011086464,
      "rewards/consensus_reward_func": 1.75,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.035330969316419214,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 4
    },
    {
      "completion_length": 5.875,
      "epoch": 0.3,
      "grad_norm": 4.957152843475342,
      "kl": 0.0,
      "learning_rate": 4.472851273490984e-07,
      "loss": 0.0,
      "reward": 3.940202221274376,
      "reward_std": 0.07331981509923935,
      "rewards/concensus_correctness_reward_func": 1.9846250116825104,
      "rewards/consensus_reward_func": 1.875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.0472959945036564,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.03328125178813934,
      "step": 6
    },
    {
      "completion_length": 4.0,
      "epoch": 0.4,
      "grad_norm": 8.158130526680907e-07,
      "kl": 2.5704503059387207e-07,
      "learning_rate": 3.867370395306068e-07,
      "loss": 0.0,
      "reward": 3.9493309408426285,
      "reward_std": 0.0,
      "rewards/concensus_correctness_reward_func": 1.923125021159649,
      "rewards/consensus_reward_func": 2.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.026205907110124826,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 8
    },
    {
      "completion_length": 4.0,
      "epoch": 0.5,
      "grad_norm": 0.00013222954294178635,
      "kl": 2.1802261471748352e-05,
      "learning_rate": 3.1137137178519977e-07,
      "loss": 0.0,
      "reward": 4.206880867481232,
      "reward_std": 0.0,
      "rewards/concensus_correctness_reward_func": 2.1730000227689743,
      "rewards/consensus_reward_func": 2.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.033880868315463886,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 10
    },
    {
      "completion_length": 24.59375,
      "epoch": 0.6,
      "grad_norm": 5.7698398450156674e-05,
      "kl": 0.0031195030896924436,
      "learning_rate": 2.2935516363191693e-07,
      "loss": 0.0,
      "reward": 3.565709076821804,
      "reward_std": 0.1642078710719943,
      "rewards/concensus_correctness_reward_func": 1.6831250041723251,
      "rewards/consensus_reward_func": 1.8125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.050552795320982113,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.01953125,
      "step": 12
    },
    {
      "completion_length": 10.09375,
      "epoch": 0.7,
      "grad_norm": 0.0004916769685223699,
      "kl": 0.0009028800996020436,
      "learning_rate": 1.4957614383675767e-07,
      "loss": 0.0,
      "reward": 3.7714153192937374,
      "reward_std": 0.0844944417476654,
      "rewards/concensus_correctness_reward_func": 1.8628750145435333,
      "rewards/consensus_reward_func": 1.875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.03354028955800459,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 14
    },
    {
      "completion_length": 4.15625,
      "epoch": 0.8,
      "grad_norm": 0.00014724327775184065,
      "kl": 0.007589483633637428,
      "learning_rate": 8.067960709356478e-08,
      "loss": 0.0,
      "reward": 3.723408281162847,
      "reward_std": 0.0002741334028542042,
      "rewards/concensus_correctness_reward_func": 1.8017500042915344,
      "rewards/consensus_reward_func": 1.875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.046658270875923336,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 16
    },
    {
      "completion_length": 12.875,
      "epoch": 0.9,
      "grad_norm": 13.287092208862305,
      "kl": 0.013432156760245562,
      "learning_rate": 3.013156219837776e-08,
      "loss": 0.0,
      "reward": 3.799438312649727,
      "reward_std": 0.12933894339948893,
      "rewards/concensus_correctness_reward_func": 1.928500022739172,
      "rewards/consensus_reward_func": 1.75,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.09040702725178562,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.030531249940395355,
      "step": 18
    },
    {
      "completion_length": 5.84375,
      "epoch": 1.0,
      "grad_norm": 0.0001615005312487483,
      "kl": 0.0014168833149597049,
      "learning_rate": 3.4096741493194193e-09,
      "loss": 0.0,
      "reward": 3.9211555421352386,
      "reward_std": 0.0394330769777298,
      "rewards/concensus_correctness_reward_func": 1.8595000132918358,
      "rewards/consensus_reward_func": 2.0,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.037311736567062326,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.02434374950826168,
      "step": 20
    },
    {
      "epoch": 1.0,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 2.6496387500907216e-06,
      "train_runtime": 958.7725,
      "train_samples_per_second": 0.334,
      "train_steps_per_second": 0.021
    }
  ],
  "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
}