trainer_state.json

{
  "best_metric": 0.9403112736054742,
  "best_model_checkpoint": "model_saves/roberta-large_spell_5k_3_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": 0.9403112736054742,
      "eval_loss": 0.42635980248451233,
      "eval_runtime": 3.377,
      "eval_samples_per_second": 1297.604,
      "eval_steps_per_second": 10.364,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.93925468755241,
      "eval_loss": 0.435956209897995,
      "eval_runtime": 3.3843,
      "eval_samples_per_second": 1294.797,
      "eval_steps_per_second": 10.342,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9385391160455729,
      "eval_loss": 0.4724823534488678,
      "eval_runtime": 3.3791,
      "eval_samples_per_second": 1296.778,
      "eval_steps_per_second": 10.358,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2897,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.938237234316126,
      "eval_loss": 0.5133076310157776,
      "eval_runtime": 3.3803,
      "eval_samples_per_second": 1296.342,
      "eval_steps_per_second": 10.354,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.387690243129344e+16,
      "train_loss": 0.2847341023274322,
      "train_runtime": 319.1771,
      "train_samples_per_second": 1612.146,
      "train_steps_per_second": 12.595
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.387690243129344e+16,
  "trial_name": null,
  "trial_params": null
}