trainer_state.json

{
  "best_metric": 0.9412672324153892,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_10k_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": 0.9412672324153892,
      "eval_loss": 0.42676153779029846,
      "eval_runtime": 4.6735,
      "eval_samples_per_second": 937.632,
      "eval_steps_per_second": 7.489,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.941121881953063,
      "eval_loss": 0.4438949525356293,
      "eval_runtime": 4.677,
      "eval_samples_per_second": 936.927,
      "eval_steps_per_second": 7.483,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9400820671071903,
      "eval_loss": 0.49137458205223083,
      "eval_runtime": 4.6845,
      "eval_samples_per_second": 935.428,
      "eval_steps_per_second": 7.471,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9398416798041123,
      "eval_loss": 0.5405893921852112,
      "eval_runtime": 4.6762,
      "eval_samples_per_second": 937.079,
      "eval_steps_per_second": 7.485,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6308703116394496e+16,
      "train_loss": 0.24491678689842794,
      "train_runtime": 475.934,
      "train_samples_per_second": 1081.158,
      "train_steps_per_second": 8.447
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6308703116394496e+16,
  "trial_name": null,
  "trial_params": null
}