trainer_state.json

{
  "best_metric": 0.9395230268674739,
  "best_model_checkpoint": "model_saves/roberta-large_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": 0.9395230268674739,
      "eval_loss": 0.4188554584980011,
      "eval_runtime": 3.3721,
      "eval_samples_per_second": 1299.476,
      "eval_steps_per_second": 10.379,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9392658683572044,
      "eval_loss": 0.4434284567832947,
      "eval_runtime": 3.3672,
      "eval_samples_per_second": 1301.362,
      "eval_steps_per_second": 10.394,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9381254262681827,
      "eval_loss": 0.46383634209632874,
      "eval_runtime": 3.3654,
      "eval_samples_per_second": 1302.078,
      "eval_steps_per_second": 10.4,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2911,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.93850557363119,
      "eval_loss": 0.5135694742202759,
      "eval_runtime": 3.3818,
      "eval_samples_per_second": 1295.76,
      "eval_steps_per_second": 10.35,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3868847622782976e+16,
      "train_loss": 0.2862282297504482,
      "train_runtime": 318.6503,
      "train_samples_per_second": 1614.811,
      "train_steps_per_second": 12.616
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3868847622782976e+16,
  "trial_name": null,
  "trial_params": null
}