{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 6.7272727272727275,
  "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": 497.15625,
      "epoch": 0.7272727272727273,
      "grad_norm": 0.3998245596885681,
      "kl": 0.0020152197321294807,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 1.1723898313939571,
      "reward_std": 1.130112662911415,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.4375,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.4880460649728775,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.24684375897049904,
      "step": 2
    },
    {
      "completion_length": 609.8571428571429,
      "epoch": 1.3636363636363638,
      "grad_norm": 0.414132297039032,
      "kl": 0.0017475592488023853,
      "learning_rate": 4.864543104251586e-07,
      "loss": 0.0,
      "reward": 2.6476131039006368,
      "reward_std": 2.906815232975142,
      "rewards/concensus_correctness_reward_func": 1.4285714285714286,
      "rewards/consensus_reward_func": 0.2857142857142857,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.21428571428571427,
      "rewards/question_recreation_reward_func": 0.6813275047710964,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.03771427060876574,
      "step": 4
    },
    {
      "completion_length": 525.4642857142857,
      "epoch": 2.0,
      "grad_norm": 0.38062697649002075,
      "kl": 0.0019122085983066686,
      "learning_rate": 4.472851273490984e-07,
      "loss": 0.0,
      "reward": 1.1200233357293266,
      "reward_std": 1.1369065046310425,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.42857142857142855,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.3834876205239977,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.017857142857142856,
      "rewards/xmlcount_reward_func": 0.2901071396523288,
      "step": 6
    },
    {
      "completion_length": 583.125,
      "epoch": 2.7272727272727275,
      "grad_norm": 2.724916458129883,
      "kl": 0.0025844083138508722,
      "learning_rate": 3.867370395306068e-07,
      "loss": 0.0,
      "reward": 1.0310951881110668,
      "reward_std": 1.2122265696525574,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.3125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.47203271416947246,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.015625,
      "rewards/xmlcount_reward_func": 0.23093750327825546,
      "step": 8
    },
    {
      "completion_length": 486.07142857142856,
      "epoch": 3.3636363636363638,
      "grad_norm": 0.4147203266620636,
      "kl": 0.0027826232669342843,
      "learning_rate": 3.1137137178519977e-07,
      "loss": 0.0,
      "reward": 2.770684165613992,
      "reward_std": 2.8159858925001964,
      "rewards/concensus_correctness_reward_func": 1.455428570508957,
      "rewards/consensus_reward_func": 0.5714285714285714,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.07142857142857142,
      "rewards/question_recreation_reward_func": 0.5599341573459762,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.112464290112257,
      "step": 10
    },
    {
      "completion_length": 547.3571428571429,
      "epoch": 4.0,
      "grad_norm": 0.25600582361221313,
      "kl": 0.0017159762127058847,
      "learning_rate": 2.2935516363191693e-07,
      "loss": 0.0,
      "reward": 2.475238765989031,
      "reward_std": 2.8428163017545427,
      "rewards/concensus_correctness_reward_func": 1.4285714285714286,
      "rewards/consensus_reward_func": 0.42857142857142855,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.46066730469465256,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.15742858286414826,
      "step": 12
    },
    {
      "completion_length": 462.21875,
      "epoch": 4.7272727272727275,
      "grad_norm": 0.6054863929748535,
      "kl": 0.001874651490652468,
      "learning_rate": 1.4957614383675767e-07,
      "loss": 0.0,
      "reward": 0.9089451059699059,
      "reward_std": 0.5815974064171314,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.5014138491824269,
      "rewards/soft_format_reward_func": 0.03125,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.31378124421462417,
      "step": 14
    },
    {
      "completion_length": 554.4285714285714,
      "epoch": 5.363636363636363,
      "grad_norm": 0.40392357110977173,
      "kl": 0.0019341647525184921,
      "learning_rate": 8.067960709356478e-08,
      "loss": 0.0,
      "reward": 0.7789905454431262,
      "reward_std": 0.8428895984377179,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.14285714285714285,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.43920483759471346,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.19692857244185039,
      "step": 16
    },
    {
      "completion_length": 517.9285714285714,
      "epoch": 6.0,
      "grad_norm": 0.39825913310050964,
      "kl": 0.002127132029272616,
      "learning_rate": 3.013156219837776e-08,
      "loss": 0.0,
      "reward": 1.6812059496130263,
      "reward_std": 2.1276232046740398,
      "rewards/concensus_correctness_reward_func": 0.7141428674970355,
      "rewards/consensus_reward_func": 0.14285714285714285,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.07142857142857142,
      "rewards/question_recreation_reward_func": 0.5451701624052865,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.2076071404985019,
      "step": 18
    },
    {
      "completion_length": 473.84375,
      "epoch": 6.7272727272727275,
      "grad_norm": 0.6048464179039001,
      "kl": 0.0023288006050279364,
      "learning_rate": 3.4096741493194193e-09,
      "loss": -0.0,
      "reward": 1.2617347463965416,
      "reward_std": 0.7358104304075823,
      "rewards/concensus_correctness_reward_func": 0.0,
      "rewards/consensus_reward_func": 0.25,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.5150159727782011,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.015625,
      "rewards/xmlcount_reward_func": 0.48109375312924385,
      "step": 20
    },
    {
      "epoch": 6.7272727272727275,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 1.3075065453449496e-06,
      "train_runtime": 1290.8252,
      "train_samples_per_second": 0.248,
      "train_steps_per_second": 0.015
    }
  ],
  "logging_steps": 2,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "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": 4,
  "trial_name": null,
  "trial_params": null
}