trainer_state.json

{
  "best_metric": 0.9391261082972753,
  "best_model_checkpoint": "model_saves/roberta-large_lemon-spell_5k_5_p3/checkpoint-804",
  "epoch": 6.0,
  "global_step": 1608,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.939120517894878,
      "eval_loss": 0.4314819872379303,
      "eval_runtime": 3.3741,
      "eval_samples_per_second": 1298.722,
      "eval_steps_per_second": 10.373,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9387180089222822,
      "eval_loss": 0.4467202425003052,
      "eval_runtime": 3.3797,
      "eval_samples_per_second": 1296.555,
      "eval_steps_per_second": 10.356,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9391261082972753,
      "eval_loss": 0.47909271717071533,
      "eval_runtime": 3.367,
      "eval_samples_per_second": 1301.443,
      "eval_steps_per_second": 10.395,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2901,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9385838392647503,
      "eval_loss": 0.5057073831558228,
      "eval_runtime": 3.3899,
      "eval_samples_per_second": 1292.649,
      "eval_steps_per_second": 10.325,
      "step": 1072
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9374433971757287,
      "eval_loss": 0.5765945315361023,
      "eval_runtime": 3.3792,
      "eval_samples_per_second": 1296.752,
      "eval_steps_per_second": 10.357,
      "step": 1340
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9384161271928353,
      "eval_loss": 0.6425763964653015,
      "eval_runtime": 3.3788,
      "eval_samples_per_second": 1296.896,
      "eval_steps_per_second": 10.359,
      "step": 1608
    },
    {
      "epoch": 6.0,
      "step": 1608,
      "total_flos": 2.0794742098886656e+16,
      "train_loss": 0.24264096502047866,
      "train_runtime": 483.4111,
      "train_samples_per_second": 1064.436,
      "train_steps_per_second": 8.316
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 2.0794742098886656e+16,
  "trial_name": null,
  "trial_params": null
}