trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "global_step": 16,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.25,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 1.0133,
      "step": 10
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.8441176470588235,
      "eval_loss": 0.5216343998908997,
      "eval_runtime": 11.8985,
      "eval_samples_per_second": 28.575,
      "eval_steps_per_second": 0.168,
      "step": 10
    },
    {
      "epoch": 2.0,
      "step": 16,
      "total_flos": 1.3938823351452303e+18,
      "train_loss": 0.8248079419136047,
      "train_runtime": 281.7391,
      "train_samples_per_second": 21.68,
      "train_steps_per_second": 0.057
    }
  ],
  "max_steps": 16,
  "num_train_epochs": 2,
  "total_flos": 1.3938823351452303e+18,
  "trial_name": null,
  "trial_params": null
}