{
  "best_metric": 0.7581556537237687,
  "best_model_checkpoint": "legal-bert-lora-no-grad/checkpoint-3858",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 5144,
  "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
    },
    {
      "epoch": 3.11,
      "learning_rate": 4.481596682218767e-05,
      "loss": 0.6117,
      "step": 2000
    },
    {
      "epoch": 3.89,
      "learning_rate": 4.3519958527734576e-05,
      "loss": 0.4942,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8109992254066615,
      "eval_f1_macro": 0.641318903734146,
      "eval_f1_micro": 0.8109992254066615,
      "eval_f1_weighted": 0.8036671817127167,
      "eval_loss": 0.6429558992385864,
      "eval_macro_fpr": 0.016902452669854486,
      "eval_macro_sensitivity": 0.6670193163886222,
      "eval_macro_specificity": 0.9858199536379567,
      "eval_precision": 0.8014184747492699,
      "eval_precision_macro": 0.6315294283267076,
      "eval_recall": 0.8109992254066615,
      "eval_recall_macro": 0.6670193163886222,
      "eval_runtime": 34.2851,
      "eval_samples_per_second": 37.655,
      "eval_steps_per_second": 4.725,
      "eval_weighted_fpr": 0.016373641121997046,
      "eval_weighted_sensitivity": 0.8109992254066615,
      "eval_weighted_specificity": 0.9763000791626885,
      "step": 2572
    },
    {
      "epoch": 4.67,
      "learning_rate": 4.22239502332815e-05,
      "loss": 0.4088,
      "step": 3000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8319132455460883,
      "eval_f1_macro": 0.7458887908906745,
      "eval_f1_micro": 0.8319132455460883,
      "eval_f1_weighted": 0.8254435175521364,
      "eval_loss": 0.7148057222366333,
      "eval_macro_fpr": 0.015014171873602745,
      "eval_macro_sensitivity": 0.7444010739541649,
      "eval_macro_specificity": 0.9871775908314393,
      "eval_precision": 0.8301095647411415,
      "eval_precision_macro": 0.795128331887366,
      "eval_recall": 0.8319132455460883,
      "eval_recall_macro": 0.7444010739541649,
      "eval_runtime": 32.6496,
      "eval_samples_per_second": 39.541,
      "eval_steps_per_second": 4.962,
      "eval_weighted_fpr": 0.014226709499770536,
      "eval_weighted_sensitivity": 0.8319132455460883,
      "eval_weighted_specificity": 0.9757506169255016,
      "step": 3215
    },
    {
      "epoch": 5.44,
      "learning_rate": 4.092794193882841e-05,
      "loss": 0.3722,
      "step": 3500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8319132455460883,
      "eval_f1_macro": 0.7581556537237687,
      "eval_f1_micro": 0.8319132455460883,
      "eval_f1_weighted": 0.8284954274383471,
      "eval_loss": 0.7203006744384766,
      "eval_macro_fpr": 0.014862611036154928,
      "eval_macro_sensitivity": 0.7604143289637364,
      "eval_macro_specificity": 0.9873237068761225,
      "eval_precision": 0.8282196199964869,
      "eval_precision_macro": 0.7594204389852343,
      "eval_recall": 0.8319132455460883,
      "eval_recall_macro": 0.7604143289637364,
      "eval_runtime": 32.5293,
      "eval_samples_per_second": 39.687,
      "eval_steps_per_second": 4.98,
      "eval_weighted_fpr": 0.014226709499770536,
      "eval_weighted_sensitivity": 0.8319132455460883,
      "eval_weighted_specificity": 0.9779423575957491,
      "step": 3858
    },
    {
      "epoch": 6.22,
      "learning_rate": 3.963193364437533e-05,
      "loss": 0.3518,
      "step": 4000
    },
    {
      "epoch": 7.0,
      "learning_rate": 3.833592534992224e-05,
      "loss": 0.3088,
      "step": 4500
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8218435321456236,
      "eval_f1_macro": 0.748587043102148,
      "eval_f1_micro": 0.8218435321456236,
      "eval_f1_weighted": 0.8202672850983224,
      "eval_loss": 0.7796419858932495,
      "eval_macro_fpr": 0.015766934174909337,
      "eval_macro_sensitivity": 0.7436163865406156,
      "eval_macro_specificity": 0.986554823310482,
      "eval_precision": 0.8283195435686048,
      "eval_precision_macro": 0.7927343403715958,
      "eval_recall": 0.8218435321456236,
      "eval_recall_macro": 0.7436163865406156,
      "eval_runtime": 33.9886,
      "eval_samples_per_second": 37.983,
      "eval_steps_per_second": 4.766,
      "eval_weighted_fpr": 0.015247944842216919,
      "eval_weighted_sensitivity": 0.8218435321456236,
      "eval_weighted_specificity": 0.9764788175116066,
      "step": 4501
    },
    {
      "epoch": 7.78,
      "learning_rate": 3.7039917055469156e-05,
      "loss": 0.245,
      "step": 5000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8187451587916343,
      "eval_f1_macro": 0.7408588571905798,
      "eval_f1_micro": 0.8187451587916342,
      "eval_f1_weighted": 0.8134714190372911,
      "eval_loss": 0.8731526136398315,
      "eval_macro_fpr": 0.016323302917471562,
      "eval_macro_sensitivity": 0.7403464639779053,
      "eval_macro_specificity": 0.98626453049897,
      "eval_precision": 0.8170725248278505,
      "eval_precision_macro": 0.7696333486311014,
      "eval_recall": 0.8187451587916343,
      "eval_recall_macro": 0.7403464639779053,
      "eval_runtime": 32.6647,
      "eval_samples_per_second": 39.523,
      "eval_steps_per_second": 4.959,
      "eval_weighted_fpr": 0.015566790846194785,
      "eval_weighted_sensitivity": 0.8187451587916343,
      "eval_weighted_specificity": 0.9752227986929156,
      "step": 5144
    }
  ],
  "logging_steps": 500,
  "max_steps": 19290,
  "num_train_epochs": 30,
  "save_steps": 500,
  "total_flos": 1.1155889738907648e+16,
  "trial_name": null,
  "trial_params": null
}