trainer_state.json

{
  "best_metric": 0.9427430986482407,
  "best_model_checkpoint": "model_saves/deberta-v3-large_spell_10k_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": 0.9427430986482407,
      "eval_loss": 0.42077574133872986,
      "eval_runtime": 4.6209,
      "eval_samples_per_second": 948.296,
      "eval_steps_per_second": 7.574,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9415747045472334,
      "eval_loss": 0.43285509943962097,
      "eval_runtime": 4.6252,
      "eval_samples_per_second": 947.412,
      "eval_steps_per_second": 7.567,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.940987712295531,
      "eval_loss": 0.45774978399276733,
      "eval_runtime": 4.6139,
      "eval_samples_per_second": 949.74,
      "eval_steps_per_second": 7.586,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2878,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9410603875266942,
      "eval_loss": 0.48844611644744873,
      "eval_runtime": 4.6276,
      "eval_samples_per_second": 946.918,
      "eval_steps_per_second": 7.563,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3800343565500416e+16,
      "train_loss": 0.2823717478495925,
      "train_runtime": 470.053,
      "train_samples_per_second": 1094.685,
      "train_steps_per_second": 8.552
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3800343565500416e+16,
  "trial_name": null,
  "trial_params": null
}