trainer_state.json

{
  "best_metric": 0.941188966781829,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-spell_5k_1_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": 0.941188966781829,
      "eval_loss": 0.4164999723434448,
      "eval_runtime": 4.6582,
      "eval_samples_per_second": 940.71,
      "eval_steps_per_second": 7.514,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9411107011482687,
      "eval_loss": 0.4404778778553009,
      "eval_runtime": 4.6516,
      "eval_samples_per_second": 942.044,
      "eval_steps_per_second": 7.524,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.940680240163687,
      "eval_loss": 0.4909224510192871,
      "eval_runtime": 4.6554,
      "eval_samples_per_second": 941.269,
      "eval_steps_per_second": 7.518,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2552,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9400932479119847,
      "eval_loss": 0.5288798809051514,
      "eval_runtime": 4.6526,
      "eval_samples_per_second": 941.846,
      "eval_steps_per_second": 7.523,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6084978873925632e+16,
      "train_loss": 0.24877668227722396,
      "train_runtime": 456.4036,
      "train_samples_per_second": 1127.423,
      "train_steps_per_second": 8.808
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6084978873925632e+16,
  "trial_name": null,
  "trial_params": null
}