checkpoint-326/trainer_state.json

{
  "best_metric": 0.3448249399662018,
  "best_model_checkpoint": "outputs/electra-nli-efl-tweeteval/checkpoint-326",
  "epoch": 1.997323135755258,
  "global_step": 326,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 9e-07,
      "loss": 0.439,
      "step": 163
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.745697074010327,
      "eval_f1": 0.732240056847258,
      "eval_loss": 0.3982622027397156,
      "eval_runtime": 8.9237,
      "eval_samples_per_second": 260.431,
      "eval_steps_per_second": 32.61,
      "step": 163
    },
    {
      "epoch": 2.0,
      "learning_rate": 8e-07,
      "loss": 0.3465,
      "step": 326
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7620481927710844,
      "eval_f1": 0.750740067157349,
      "eval_loss": 0.3448249399662018,
      "eval_runtime": 9.0059,
      "eval_samples_per_second": 258.053,
      "eval_steps_per_second": 32.312,
      "step": 326
    }
  ],
  "max_steps": 1630,
  "num_train_epochs": 10,
  "total_flos": 6097705758760512.0,
  "trial_name": null,
  "trial_params": null
}