training_checkpoints/checkpoint-1929/trainer_state.json

{
  "best_metric": 0.7369957890904685,
  "best_model_checkpoint": "legal-bert-base-uncased/checkpoint-1929",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1929,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.78,
      "grad_norm": 6.735908031463623,
      "learning_rate": 4.742353551062727e-05,
      "loss": 1.1096,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7978311386522076,
      "eval_f1_macro": 0.613418966848975,
      "eval_f1_micro": 0.7978311386522076,
      "eval_f1_weighted": 0.7839880886224406,
      "eval_loss": 0.6747979521751404,
      "eval_macro_fpr": 0.018830502815301575,
      "eval_macro_sensitivity": 0.6339578261148517,
      "eval_macro_specificity": 0.9845345883924012,
      "eval_precision": 0.7854860044786733,
      "eval_precision_macro": 0.6239440349375213,
      "eval_recall": 0.7978311386522076,
      "eval_recall_macro": 0.6339578261148517,
      "eval_runtime": 30.9921,
      "eval_samples_per_second": 41.656,
      "eval_steps_per_second": 5.227,
      "eval_weighted_fpr": 0.017778080512226686,
      "eval_weighted_sensitivity": 0.7978311386522076,
      "eval_weighted_specificity": 0.9701876872338092,
      "step": 643
    },
    {
      "epoch": 1.56,
      "grad_norm": 6.824150562286377,
      "learning_rate": 4.48315189217211e-05,
      "loss": 0.6187,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8109992254066615,
      "eval_f1_macro": 0.7268228216676939,
      "eval_f1_micro": 0.8109992254066615,
      "eval_f1_weighted": 0.8089875323563979,
      "eval_loss": 0.6449354887008667,
      "eval_macro_fpr": 0.016938802987367406,
      "eval_macro_sensitivity": 0.7327307314578203,
      "eval_macro_specificity": 0.9857686909763926,
      "eval_precision": 0.8196084470842065,
      "eval_precision_macro": 0.7805936449528967,
      "eval_recall": 0.8109992254066615,
      "eval_recall_macro": 0.7327307314578203,
      "eval_runtime": 58.3071,
      "eval_samples_per_second": 22.141,
      "eval_steps_per_second": 2.778,
      "eval_weighted_fpr": 0.016373641121997046,
      "eval_weighted_sensitivity": 0.8109992254066615,
      "eval_weighted_specificity": 0.9755311392392266,
      "step": 1286
    },
    {
      "epoch": 2.33,
      "grad_norm": 20.75279998779297,
      "learning_rate": 4.223950233281493e-05,
      "loss": 0.4747,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.814872192099148,
      "eval_f1_macro": 0.7369957890904685,
      "eval_f1_micro": 0.814872192099148,
      "eval_f1_weighted": 0.8124596447054482,
      "eval_loss": 0.8151384592056274,
      "eval_macro_fpr": 0.016571707879599946,
      "eval_macro_sensitivity": 0.739039840019417,
      "eval_macro_specificity": 0.9860629698102752,
      "eval_precision": 0.8192276160546627,
      "eval_precision_macro": 0.7659183928188651,
      "eval_recall": 0.814872192099148,
      "eval_recall_macro": 0.739039840019417,
      "eval_runtime": 55.9094,
      "eval_samples_per_second": 23.091,
      "eval_steps_per_second": 2.898,
      "eval_weighted_fpr": 0.015968463954032203,
      "eval_weighted_sensitivity": 0.814872192099148,
      "eval_weighted_specificity": 0.9760723550549795,
      "step": 1929
    }
  ],
  "logging_steps": 500,
  "max_steps": 9645,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 4056067183343616.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}