trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0,
  "eval_steps": 100,
  "global_step": 0,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 462.55078125,
      "epoch": 0,
      "kl": 0.0,
      "reward": 0.453125,
      "reward_std": 0.46982569620013237,
      "rewards/accuracy_reward": 0.08984375,
      "rewards/format_reward": 0.36328125,
      "step": 0,
      "total_flos": 0.0,
      "train_loss": -42.93539375066757,
      "train_runtime": 225.8365,
      "train_samples_per_second": 0.177,
      "train_steps_per_second": 0.009
    }
  ],
  "logging_steps": 5,
  "max_steps": 2,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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": 4,
  "trial_name": null,
  "trial_params": null
}