trainer_state.json

{
  "best_metric": 0.9415858853520276,
  "best_model_checkpoint": "model_saves/deberta-large_spell_5k_2_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": 0.9415858853520276,
      "eval_loss": 0.4140821695327759,
      "eval_runtime": 4.6367,
      "eval_samples_per_second": 945.078,
      "eval_steps_per_second": 7.549,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9412113283914176,
      "eval_loss": 0.43673956394195557,
      "eval_runtime": 4.6514,
      "eval_samples_per_second": 942.088,
      "eval_steps_per_second": 7.525,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9400261630832187,
      "eval_loss": 0.4806632399559021,
      "eval_runtime": 4.6427,
      "eval_samples_per_second": 943.853,
      "eval_steps_per_second": 7.539,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.255,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9397745949753463,
      "eval_loss": 0.5354779362678528,
      "eval_runtime": 4.6445,
      "eval_samples_per_second": 943.488,
      "eval_steps_per_second": 7.536,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6098442656874496e+16,
      "train_loss": 0.2488155276028078,
      "train_runtime": 461.8978,
      "train_samples_per_second": 1114.013,
      "train_steps_per_second": 8.703
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6098442656874496e+16,
  "trial_name": null,
  "trial_params": null
}