trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 63,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.57,
      "learning_rate": 4.761904761904762e-05,
      "loss": 0.36,
      "step": 33
    },
    {
      "epoch": 1.57,
      "eval_accuracy": 0.9124251497005988,
      "eval_f1": 0.9341586944288126,
      "eval_loss": 0.22648876905441284,
      "eval_precision": 0.9315375982042648,
      "eval_recall": 0.9367945823927766,
      "eval_roc_auc": 0.9006195134186106,
      "eval_runtime": 1.0671,
      "eval_samples_per_second": 1252.042,
      "eval_steps_per_second": 5.623,
      "step": 33
    },
    {
      "epoch": 3.0,
      "step": 63,
      "total_flos": 1061461265430528.0,
      "train_loss": 0.2692271187191918,
      "train_runtime": 63.0344,
      "train_samples_per_second": 254.242,
      "train_steps_per_second": 0.999
    }
  ],
  "max_steps": 63,
  "num_train_epochs": 3,
  "total_flos": 1061461265430528.0,
  "trial_name": null,
  "trial_params": null
}