trainer_state.json

{
  "best_metric": 0.9417759590335313,
  "best_model_checkpoint": "model_saves/deberta-v3-large_lemon_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": 0.9417759590335313,
      "eval_loss": 0.41643938422203064,
      "eval_runtime": 4.5349,
      "eval_samples_per_second": 966.292,
      "eval_steps_per_second": 7.718,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9407529153948502,
      "eval_loss": 0.43404144048690796,
      "eval_runtime": 4.527,
      "eval_samples_per_second": 967.969,
      "eval_steps_per_second": 7.731,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9406019745301267,
      "eval_loss": 0.4636657238006592,
      "eval_runtime": 4.5338,
      "eval_samples_per_second": 966.508,
      "eval_steps_per_second": 7.72,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2651,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9396348349154172,
      "eval_loss": 0.5008194446563721,
      "eval_runtime": 4.5168,
      "eval_samples_per_second": 970.152,
      "eval_steps_per_second": 7.749,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3570052787798016e+16,
      "train_loss": 0.2598130026860024,
      "train_runtime": 474.6409,
      "train_samples_per_second": 1084.104,
      "train_steps_per_second": 8.47
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3570052787798016e+16,
  "trial_name": null,
  "trial_params": null
}