{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8,
  "eval_steps": 500,
  "global_step": 4,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 994.5480270385742,
      "epoch": 0.2,
      "grad_norm": 150.32339477539062,
      "kl": 0.0,
      "learning_rate": 3e-06,
      "loss": -0.0199,
      "reward": 0.7332589626312256,
      "reward_std": 0.41786693036556244,
      "rewards/accuracy_reward": 0.06808035937137902,
      "rewards/format_reward": 0.6651786044239998,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 989.4696756998698,
      "epoch": 0.8,
      "kl": 0.9493815104166666,
      "reward": 0.7220982511838278,
      "reward_std": 0.3837023004889488,
      "rewards/accuracy_reward": 0.049851192510686815,
      "rewards/format_reward": 0.6722470571597418,
      "step": 4,
      "total_flos": 0.0,
      "train_loss": 0.013942703604698181,
      "train_runtime": 1365.9206,
      "train_samples_per_second": 0.462,
      "train_steps_per_second": 0.004
    }
  ],
  "logging_steps": 5,
  "max_steps": 5,
  "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
}