trainer_state.json

{
  "best_metric": 0.9423685416876306,
  "best_model_checkpoint": "model_saves/deberta-v3-large_spell_10k_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": 0.9423685416876306,
      "eval_loss": 0.4189268946647644,
      "eval_runtime": 4.6367,
      "eval_samples_per_second": 945.063,
      "eval_steps_per_second": 7.548,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9422902760540703,
      "eval_loss": 0.4353422224521637,
      "eval_runtime": 4.6212,
      "eval_samples_per_second": 948.241,
      "eval_steps_per_second": 7.574,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9416361989736021,
      "eval_loss": 0.4561524987220764,
      "eval_runtime": 4.6186,
      "eval_samples_per_second": 948.766,
      "eval_steps_per_second": 7.578,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2882,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9407640961996444,
      "eval_loss": 0.486333429813385,
      "eval_runtime": 4.6237,
      "eval_samples_per_second": 947.724,
      "eval_steps_per_second": 7.57,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.377969893277696e+16,
      "train_loss": 0.2827758557760893,
      "train_runtime": 473.2316,
      "train_samples_per_second": 1087.332,
      "train_steps_per_second": 8.495
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.377969893277696e+16,
  "trial_name": null,
  "trial_params": null
}