trainer_state.json

{
  "best_metric": 0.9421840584085243,
  "best_model_checkpoint": "model_saves/deberta-v3-large_spell_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": 0.9421840584085243,
      "eval_loss": 0.40784236788749695,
      "eval_runtime": 4.5824,
      "eval_samples_per_second": 956.27,
      "eval_steps_per_second": 7.638,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9419828039222263,
      "eval_loss": 0.4205301105976105,
      "eval_runtime": 4.6312,
      "eval_samples_per_second": 946.192,
      "eval_steps_per_second": 7.557,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9404901664821834,
      "eval_loss": 0.44686171412467957,
      "eval_runtime": 4.5721,
      "eval_samples_per_second": 958.418,
      "eval_steps_per_second": 7.655,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2926,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9400429342904102,
      "eval_loss": 0.4810025990009308,
      "eval_runtime": 4.5676,
      "eval_samples_per_second": 959.359,
      "eval_steps_per_second": 7.663,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3574302590828544e+16,
      "train_loss": 0.2871383117206061,
      "train_runtime": 468.9433,
      "train_samples_per_second": 1097.276,
      "train_steps_per_second": 8.572
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3574302590828544e+16,
  "trial_name": null,
  "trial_params": null
}