trainer_state.json

{
  "best_metric": 0.9408367714308076,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-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": 0.9408367714308076,
      "eval_loss": 0.43003103137016296,
      "eval_runtime": 4.6392,
      "eval_samples_per_second": 944.567,
      "eval_steps_per_second": 7.544,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9396571965250059,
      "eval_loss": 0.4691629409790039,
      "eval_runtime": 4.6421,
      "eval_samples_per_second": 943.961,
      "eval_steps_per_second": 7.54,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9393441339907647,
      "eval_loss": 0.5035601258277893,
      "eval_runtime": 4.6505,
      "eval_samples_per_second": 942.268,
      "eval_steps_per_second": 7.526,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2201,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.94000380147363,
      "eval_loss": 0.5705075263977051,
      "eval_runtime": 4.656,
      "eval_samples_per_second": 941.158,
      "eval_steps_per_second": 7.517,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6105037109395456e+16,
      "train_loss": 0.21437349604136907,
      "train_runtime": 458.6729,
      "train_samples_per_second": 1121.845,
      "train_steps_per_second": 8.764
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6105037109395456e+16,
  "trial_name": null,
  "trial_params": null
}