{
  "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": 237.1875,
      "epoch": 0.1,
      "grad_norm": 8.398701667785645,
      "kl": 0.0,
      "learning_rate": 4.965903258506806e-07,
      "loss": 0.0,
      "reward": 0.09728920509223826,
      "reward_std": 0.05032656194816809,
      "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": 0.09338295520865358,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.00390625,
      "step": 2
    },
    {
      "completion_length": 254.8125,
      "epoch": 0.2,
      "grad_norm": 4.527547359466553,
      "kl": 0.0005934742662248027,
      "learning_rate": 4.698684378016222e-07,
      "loss": 0.0,
      "reward": 0.10370710026472807,
      "reward_std": 0.056229994814202655,
      "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": 0.10933210048824549,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.005625000223517418,
      "step": 4
    },
    {
      "completion_length": 198.375,
      "epoch": 0.3,
      "grad_norm": 16.620403289794922,
      "kl": 0.011071403857840778,
      "learning_rate": 4.193203929064353e-07,
      "loss": 0.0,
      "reward": 0.17157541977940127,
      "reward_std": 0.12756577530672075,
      "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": 0.1181379213812761,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.009062499972060323,
      "step": 6
    },
    {
      "completion_length": 151.34375,
      "epoch": 0.4,
      "grad_norm": 5.440056324005127,
      "kl": 0.0035489053407218307,
      "learning_rate": 3.5042385616324236e-07,
      "loss": 0.0,
      "reward": 0.12494469387456775,
      "reward_std": 0.02955692503019236,
      "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": 0.12494469387456775,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 8
    },
    {
      "completion_length": 159.4375,
      "epoch": 0.5,
      "grad_norm": 17.10235595703125,
      "kl": 0.019251713070843834,
      "learning_rate": 2.706448363680831e-07,
      "loss": 0.0,
      "reward": 0.2264014035463333,
      "reward_std": 0.1991720616351813,
      "rewards/concensus_correctness_reward_func": 0.020625000819563866,
      "rewards/consensus_reward_func": 0.0625,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.12343264813534915,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.019843749701976776,
      "step": 10
    },
    {
      "completion_length": 161.6875,
      "epoch": 0.6,
      "grad_norm": 13.440507888793945,
      "kl": 0.01697970710665686,
      "learning_rate": 1.886286282148002e-07,
      "loss": 0.0,
      "reward": 0.30120708944741637,
      "reward_std": 0.30583049908454996,
      "rewards/concensus_correctness_reward_func": 0.0364375002682209,
      "rewards/consensus_reward_func": 0.125,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.14223835291340947,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.0024687498807907104,
      "step": 12
    },
    {
      "completion_length": 142.25,
      "epoch": 0.7,
      "grad_norm": 24.540307998657227,
      "kl": 0.0875492471823236,
      "learning_rate": 1.1326296046939333e-07,
      "loss": 0.0001,
      "reward": 0.5317345891380683,
      "reward_std": 0.23181627465237398,
      "rewards/concensus_correctness_reward_func": 0.04262499883770943,
      "rewards/consensus_reward_func": 0.1875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.1875,
      "rewards/question_recreation_reward_func": 0.0995783495309297,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.014531250111758709,
      "step": 14
    },
    {
      "completion_length": 130.9375,
      "epoch": 0.8,
      "grad_norm": 24.440187454223633,
      "kl": 0.062173418860766105,
      "learning_rate": 5.271487265090163e-08,
      "loss": 0.0001,
      "reward": 0.5183445231523365,
      "reward_std": 0.6032340360834496,
      "rewards/concensus_correctness_reward_func": 0.13956249877810478,
      "rewards/consensus_reward_func": 0.1875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0625,
      "rewards/question_recreation_reward_func": 0.14553201623493806,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": -0.016750000417232513,
      "step": 16
    },
    {
      "completion_length": 118.03125,
      "epoch": 0.9,
      "grad_norm": 27.92249870300293,
      "kl": 0.03141187228175113,
      "learning_rate": 1.3545689574841341e-08,
      "loss": 0.0,
      "reward": 0.3571384238312021,
      "reward_std": 0.3553317471523769,
      "rewards/concensus_correctness_reward_func": 0.019187500700354576,
      "rewards/consensus_reward_func": 0.1875,
      "rewards/cumulative_reward_2": 0.0,
      "rewards/final_correctness_reward_func": 0.0,
      "rewards/question_recreation_reward_func": 0.15045091963838786,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 18
    },
    {
      "completion_length": 104.75,
      "epoch": 1.0,
      "grad_norm": 5.20966100692749,
      "kl": 0.07331209685071371,
      "learning_rate": 0.0,
      "loss": 0.0001,
      "reward": 0.12898236550972797,
      "reward_std": 0.06363634395529516,
      "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": 0.12898236550972797,
      "rewards/soft_format_reward_func": 0.0,
      "rewards/strict_format_reward_func": 0.0,
      "rewards/xmlcount_reward_func": 0.0,
      "step": 20
    },
    {
      "epoch": 1.0,
      "step": 20,
      "total_flos": 0.0,
      "train_loss": 3.0612738328272826e-05,
      "train_runtime": 915.7967,
      "train_samples_per_second": 0.349,
      "train_steps_per_second": 0.022
    }
  ],
  "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
}