{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9893390191897654,
  "eval_steps": 100,
  "global_step": 58,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 598.0323905944824,
      "epoch": 0.017057569296375266,
      "grad_norm": 0.345001757144928,
      "kl": 0.0,
      "learning_rate": 5e-07,
      "loss": 0.0236,
      "reward": 0.6529018133878708,
      "reward_std": 0.33456145599484444,
      "rewards/accuracy_reward": 0.6529018133878708,
      "rewards/format_reward": 0.0,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 598.3013668060303,
      "epoch": 0.08528784648187633,
      "grad_norm": 0.41064807772636414,
      "kl": 0.00019729137420654297,
      "learning_rate": 2.5e-06,
      "loss": 0.0203,
      "reward": 0.6194196660071611,
      "reward_std": 0.3350280746817589,
      "rewards/accuracy_reward": 0.6194196660071611,
      "rewards/format_reward": 0.0,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 613.2944435119629,
      "epoch": 0.17057569296375266,
      "grad_norm": 0.163320854306221,
      "kl": 0.004157018661499023,
      "learning_rate": 2.956412726139078e-06,
      "loss": 0.0591,
      "reward": 0.6962053909897804,
      "reward_std": 0.28323794081807135,
      "rewards/accuracy_reward": 0.6962053909897804,
      "rewards/format_reward": 0.0,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 602.157389831543,
      "epoch": 0.255863539445629,
      "grad_norm": 0.3424699008464813,
      "kl": 0.002886962890625,
      "learning_rate": 2.7836719084521715e-06,
      "loss": 0.073,
      "reward": 0.7660714626312256,
      "reward_std": 0.20958704613149165,
      "rewards/accuracy_reward": 0.7660714626312256,
      "rewards/format_reward": 0.0,
      "step": 15
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 584.3727920532226,
      "epoch": 0.3411513859275053,
      "grad_norm": 0.13450783491134644,
      "kl": 0.003928756713867188,
      "learning_rate": 2.4946839873611927e-06,
      "loss": 0.0495,
      "reward": 0.7785714611411094,
      "reward_std": 0.18694723900407553,
      "rewards/accuracy_reward": 0.7785714611411094,
      "rewards/format_reward": 0.0,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 593.649806213379,
      "epoch": 0.42643923240938164,
      "grad_norm": 0.18525253236293793,
      "kl": 0.004283523559570313,
      "learning_rate": 2.1156192081791355e-06,
      "loss": 0.0406,
      "reward": 0.7669643223285675,
      "reward_std": 0.17252950482070445,
      "rewards/accuracy_reward": 0.7669643223285675,
      "rewards/format_reward": 0.0,
      "step": 25
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 595.1203407287597,
      "epoch": 0.511727078891258,
      "grad_norm": 0.16254669427871704,
      "kl": 0.0033504486083984373,
      "learning_rate": 1.6808050203829845e-06,
      "loss": 0.0357,
      "reward": 0.7613839641213417,
      "reward_std": 0.1727013595402241,
      "rewards/accuracy_reward": 0.7613839641213417,
      "rewards/format_reward": 0.0,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 590.8337310791015,
      "epoch": 0.5970149253731343,
      "grad_norm": 0.25264808535575867,
      "kl": 0.006267547607421875,
      "learning_rate": 1.2296174432791415e-06,
      "loss": 0.036,
      "reward": 0.7473214708268643,
      "reward_std": 0.17480504335835576,
      "rewards/accuracy_reward": 0.7473214708268643,
      "rewards/format_reward": 0.0,
      "step": 35
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 576.3962310791015,
      "epoch": 0.6823027718550106,
      "grad_norm": 0.11616054177284241,
      "kl": 0.006030654907226563,
      "learning_rate": 8.029152419343472e-07,
      "loss": 0.0366,
      "reward": 0.7694196775555611,
      "reward_std": 0.17429448990151286,
      "rewards/accuracy_reward": 0.7694196775555611,
      "rewards/format_reward": 0.0,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 598.3955627441406,
      "epoch": 0.767590618336887,
      "grad_norm": 0.5602137446403503,
      "kl": 0.00694427490234375,
      "learning_rate": 4.3933982822017883e-07,
      "loss": 0.0348,
      "reward": 0.7482143223285675,
      "reward_std": 0.19791895523667336,
      "rewards/accuracy_reward": 0.7482143223285675,
      "rewards/format_reward": 0.0,
      "step": 45
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 596.3730171203613,
      "epoch": 0.8528784648187633,
      "grad_norm": 41.710044860839844,
      "kl": 0.02411231994628906,
      "learning_rate": 1.718159615201853e-07,
      "loss": 0.0338,
      "reward": 0.7491071775555611,
      "reward_std": 0.18286419473588467,
      "rewards/accuracy_reward": 0.7491071775555611,
      "rewards/format_reward": 0.0,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 589.5062797546386,
      "epoch": 0.9381663113006397,
      "grad_norm": 0.1343313455581665,
      "kl": 0.005272674560546875,
      "learning_rate": 2.4570139579284723e-08,
      "loss": 0.0328,
      "reward": 0.7848214715719223,
      "reward_std": 0.19012680910527707,
      "rewards/accuracy_reward": 0.7848214715719223,
      "rewards/format_reward": 0.0,
      "step": 55
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 582.7546895345052,
      "epoch": 0.9893390191897654,
      "kl": 0.0041383107503255205,
      "reward": 0.761904795964559,
      "reward_std": 0.1910448515166839,
      "rewards/accuracy_reward": 0.761904795964559,
      "rewards/format_reward": 0.0,
      "step": 58,
      "total_flos": 0.0,
      "train_loss": 0.03878247848679793,
      "train_runtime": 12767.7366,
      "train_samples_per_second": 0.587,
      "train_steps_per_second": 0.005
    }
  ],
  "logging_steps": 5,
  "max_steps": 58,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}