trainer_state.json

{
  "best_metric": 0.9396180637082258,
  "best_model_checkpoint": "model_saves/roberta-large_lemon_5k_3_p3/checkpoint-268",
  "epoch": 4.0,
  "global_step": 1072,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9396180637082258,
      "eval_loss": 0.430489182472229,
      "eval_runtime": 3.3882,
      "eval_samples_per_second": 1293.329,
      "eval_steps_per_second": 10.33,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9392658683572044,
      "eval_loss": 0.4401320219039917,
      "eval_runtime": 3.3864,
      "eval_samples_per_second": 1293.988,
      "eval_steps_per_second": 10.335,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9387459609342681,
      "eval_loss": 0.47696542739868164,
      "eval_runtime": 3.3779,
      "eval_samples_per_second": 1297.24,
      "eval_steps_per_second": 10.361,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2914,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9392770491619987,
      "eval_loss": 0.5186468362808228,
      "eval_runtime": 3.3837,
      "eval_samples_per_second": 1295.029,
      "eval_steps_per_second": 10.344,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3863634941771776e+16,
      "train_loss": 0.28617537199561277,
      "train_runtime": 318.5269,
      "train_samples_per_second": 1615.437,
      "train_steps_per_second": 12.621
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3863634941771776e+16,
  "trial_name": null,
  "trial_params": null
}