trainer_state.json

{
  "best_metric": 0.942105792774964,
  "best_model_checkpoint": "model_saves/deberta-v3-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": 0.942105792774964,
      "eval_loss": 0.4105583429336548,
      "eval_runtime": 4.5175,
      "eval_samples_per_second": 969.997,
      "eval_steps_per_second": 7.748,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9417703686311341,
      "eval_loss": 0.4260367155075073,
      "eval_runtime": 4.5375,
      "eval_samples_per_second": 965.739,
      "eval_steps_per_second": 7.714,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9413846308657298,
      "eval_loss": 0.4569767713546753,
      "eval_runtime": 4.5148,
      "eval_samples_per_second": 970.583,
      "eval_steps_per_second": 7.752,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2933,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9402218271671194,
      "eval_loss": 0.48914003372192383,
      "eval_runtime": 4.5137,
      "eval_samples_per_second": 970.819,
      "eval_steps_per_second": 7.754,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3559663664562176e+16,
      "train_loss": 0.28773752077301934,
      "train_runtime": 464.8739,
      "train_samples_per_second": 1106.881,
      "train_steps_per_second": 8.648
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3559663664562176e+16,
  "trial_name": null,
  "trial_params": null
}