run-1/checkpoint-500/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.5723270440251573,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6690322580645162,
      "eval_loss": 0.42948082089424133,
      "eval_runtime": 5.393,
      "eval_samples_per_second": 574.82,
      "eval_steps_per_second": 12.053,
      "step": 318
    },
    {
      "epoch": 1.5723270440251573,
      "grad_norm": 0.8646999001502991,
      "learning_rate": 1.371069182389937e-05,
      "loss": 0.6591,
      "step": 500
    }
  ],
  "logging_steps": 500,
  "max_steps": 1590,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 130072209152340.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.4868927644085208,
    "num_train_epochs": 5,
    "temperature": 2
  }
}