trainer_state.json

{
  "best_metric": 0.9403671776294458,
  "best_model_checkpoint": "model_saves/roberta-large_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": 0.9403671776294458,
      "eval_loss": 0.4260515868663788,
      "eval_runtime": 3.3609,
      "eval_samples_per_second": 1303.814,
      "eval_steps_per_second": 10.414,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.940238598374311,
      "eval_loss": 0.44465357065200806,
      "eval_runtime": 3.3583,
      "eval_samples_per_second": 1304.83,
      "eval_steps_per_second": 10.422,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9395062556602825,
      "eval_loss": 0.4736112952232361,
      "eval_runtime": 3.3606,
      "eval_samples_per_second": 1303.928,
      "eval_steps_per_second": 10.415,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2895,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9381701494873601,
      "eval_loss": 0.5212493538856506,
      "eval_runtime": 3.3545,
      "eval_samples_per_second": 1306.29,
      "eval_steps_per_second": 10.434,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3880930238201856e+16,
      "train_loss": 0.28438915957265826,
      "train_runtime": 321.1747,
      "train_samples_per_second": 1602.119,
      "train_steps_per_second": 12.517
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3880930238201856e+16,
  "trial_name": null,
  "trial_params": null
}