checkpoint-652/trainer_state.json

{
  "best_metric": 0.3068828284740448,
  "best_model_checkpoint": "outputs/electra-nli-efl-tweeteval/checkpoint-652",
  "epoch": 3.9973231357552583,
  "global_step": 652,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 9.499999999999999e-07,
      "loss": 0.4384,
      "step": 163
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7444061962134251,
      "eval_f1": 0.7308261375858633,
      "eval_loss": 0.39615127444267273,
      "eval_runtime": 9.0599,
      "eval_samples_per_second": 256.514,
      "eval_steps_per_second": 32.119,
      "step": 163
    },
    {
      "epoch": 2.0,
      "learning_rate": 9e-07,
      "loss": 0.3447,
      "step": 326
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.76592082616179,
      "eval_f1": 0.7552159046464709,
      "eval_loss": 0.3409559428691864,
      "eval_runtime": 9.4378,
      "eval_samples_per_second": 246.244,
      "eval_steps_per_second": 30.833,
      "step": 326
    },
    {
      "epoch": 3.0,
      "learning_rate": 8.499999999999999e-07,
      "loss": 0.3057,
      "step": 489
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7749569707401033,
      "eval_f1": 0.768808341108185,
      "eval_loss": 0.32338443398475647,
      "eval_runtime": 9.0418,
      "eval_samples_per_second": 257.028,
      "eval_steps_per_second": 32.184,
      "step": 489
    },
    {
      "epoch": 4.0,
      "learning_rate": 8e-07,
      "loss": 0.287,
      "step": 652
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7857142857142857,
      "eval_f1": 0.7778970154753132,
      "eval_loss": 0.3068828284740448,
      "eval_runtime": 9.3662,
      "eval_samples_per_second": 248.128,
      "eval_steps_per_second": 31.069,
      "step": 652
    }
  ],
  "max_steps": 3260,
  "num_train_epochs": 20,
  "total_flos": 1.2184053565654464e+16,
  "trial_name": null,
  "trial_params": null
}