training_checkpoints/checkpoint-1929/trainer_state.json

{
  "best_metric": 0.5439483491438576,
  "best_model_checkpoint": "legal-bert-lora-no-grad/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,
      "learning_rate": 4.870399170554692e-05,
      "loss": 1.5473,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7172734314484895,
      "eval_f1_macro": 0.3920432121625908,
      "eval_f1_micro": 0.7172734314484894,
      "eval_f1_weighted": 0.6868869808248764,
      "eval_loss": 0.8484711050987244,
      "eval_macro_fpr": 0.027836504093424867,
      "eval_macro_sensitivity": 0.44215864388540455,
      "eval_macro_specificity": 0.9788744549045134,
      "eval_precision": 0.6896842532804647,
      "eval_precision_macro": 0.3892290687521846,
      "eval_recall": 0.7172734314484895,
      "eval_recall_macro": 0.44215864388540455,
      "eval_runtime": 32.4222,
      "eval_samples_per_second": 39.818,
      "eval_steps_per_second": 4.997,
      "eval_weighted_fpr": 0.027383899767424412,
      "eval_weighted_sensitivity": 0.7172734314484895,
      "eval_weighted_specificity": 0.9658433921192108,
      "step": 643
    },
    {
      "epoch": 1.56,
      "learning_rate": 4.740798341109383e-05,
      "loss": 0.7816,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7544539116963594,
      "eval_f1_macro": 0.47513531679703136,
      "eval_f1_micro": 0.7544539116963594,
      "eval_f1_weighted": 0.7304002619527014,
      "eval_loss": 0.7113391757011414,
      "eval_macro_fpr": 0.023051270843305598,
      "eval_macro_sensitivity": 0.5054156753128263,
      "eval_macro_specificity": 0.9817050378554314,
      "eval_precision": 0.7491876907285084,
      "eval_precision_macro": 0.5281699878111706,
      "eval_recall": 0.7544539116963594,
      "eval_recall_macro": 0.5054156753128263,
      "eval_runtime": 33.6622,
      "eval_samples_per_second": 38.352,
      "eval_steps_per_second": 4.813,
      "eval_weighted_fpr": 0.02271912850283093,
      "eval_weighted_sensitivity": 0.7544539116963594,
      "eval_weighted_specificity": 0.9711216561351127,
      "step": 1286
    },
    {
      "epoch": 2.33,
      "learning_rate": 4.6111975116640746e-05,
      "loss": 0.6956,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7986057319907048,
      "eval_f1_macro": 0.5439483491438576,
      "eval_f1_micro": 0.7986057319907048,
      "eval_f1_weighted": 0.7833451448204543,
      "eval_loss": 0.6460158824920654,
      "eval_macro_fpr": 0.01844352926124183,
      "eval_macro_sensitivity": 0.5700868724489073,
      "eval_macro_specificity": 0.9848453217277398,
      "eval_precision": 0.7793241511142787,
      "eval_precision_macro": 0.5435553848320287,
      "eval_recall": 0.7986057319907048,
      "eval_recall_macro": 0.5700868724489073,
      "eval_runtime": 33.205,
      "eval_samples_per_second": 38.88,
      "eval_steps_per_second": 4.879,
      "eval_weighted_fpr": 0.01769429699196951,
      "eval_weighted_sensitivity": 0.7986057319907048,
      "eval_weighted_specificity": 0.974074093925392,
      "step": 1929
    }
  ],
  "logging_steps": 500,
  "max_steps": 19290,
  "num_train_epochs": 30,
  "save_steps": 500,
  "total_flos": 4183458652090368.0,
  "trial_name": null,
  "trial_params": null
}