trainer_state.json

{
  "best_metric": 0.9395901116962398,
  "best_model_checkpoint": "model_saves/roberta-large_spell_5k_6_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": 0.9395901116962398,
      "eval_loss": 0.4252014756202698,
      "eval_runtime": 3.3489,
      "eval_samples_per_second": 1308.481,
      "eval_steps_per_second": 10.451,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9391149274924809,
      "eval_loss": 0.4422414004802704,
      "eval_runtime": 3.3533,
      "eval_samples_per_second": 1306.755,
      "eval_steps_per_second": 10.437,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9390590234685092,
      "eval_loss": 0.48049771785736084,
      "eval_runtime": 3.3583,
      "eval_samples_per_second": 1304.832,
      "eval_steps_per_second": 10.422,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2889,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9383210903520836,
      "eval_loss": 0.5132158994674683,
      "eval_runtime": 3.3473,
      "eval_samples_per_second": 1309.109,
      "eval_steps_per_second": 10.456,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.387974583386112e+16,
      "train_loss": 0.2840782786483195,
      "train_runtime": 342.691,
      "train_samples_per_second": 1501.527,
      "train_steps_per_second": 11.731
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.387974583386112e+16,
  "trial_name": null,
  "trial_params": null
}