run-16/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": 0.7283870967741936,
      "eval_loss": 3.1895267963409424,
      "eval_runtime": 2.407,
      "eval_samples_per_second": 1287.91,
      "eval_steps_per_second": 27.005,
      "step": 318
    },
    {
      "epoch": 1.57,
      "learning_rate": 1.606918238993711e-05,
      "loss": 3.7135,
      "step": 500
    }
  ],
  "logging_steps": 500,
  "max_steps": 2544,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 188679284472168.0,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.5656189676317626,
    "num_train_epochs": 8,
    "temperature": 9
  }
}