training_checkpoints/checkpoint-1120/trainer_state.json

{
  "best_metric": 0.5892045454545454,
  "best_model_checkpoint": "case-analysis-roberta-base/checkpoint-1120",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 1120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7527839643652561,
      "eval_f1_macro": 0.5389063030773956,
      "eval_f1_micro": 0.7527839643652561,
      "eval_f1_weighted": 0.7300520074128541,
      "eval_loss": 0.8771109580993652,
      "eval_macro_fpr": 0.13144039552419665,
      "eval_macro_sensitivity": 0.5401895401895401,
      "eval_macro_specificity": 0.8833499134100214,
      "eval_precision": 0.7119932520850578,
      "eval_precision_macro": 0.5403637213518124,
      "eval_recall": 0.7527839643652561,
      "eval_recall_macro": 0.5401895401895401,
      "eval_runtime": 6.9264,
      "eval_samples_per_second": 64.825,
      "eval_steps_per_second": 8.229,
      "eval_weighted_fpr": 0.09866666666666667,
      "eval_weighted_sensitivity": 0.7527839643652561,
      "eval_weighted_specificity": 0.7806156892748298,
      "step": 224
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7728285077951003,
      "eval_f1_macro": 0.5712387529890991,
      "eval_f1_micro": 0.7728285077951002,
      "eval_f1_weighted": 0.7555182045354769,
      "eval_loss": 0.7935857176780701,
      "eval_macro_fpr": 0.10804320901264429,
      "eval_macro_sensitivity": 0.59374376286141,
      "eval_macro_specificity": 0.9046495531497725,
      "eval_precision": 0.7420199260104934,
      "eval_precision_macro": 0.5529178338001868,
      "eval_recall": 0.7728285077951003,
      "eval_recall_macro": 0.59374376286141,
      "eval_runtime": 33.5269,
      "eval_samples_per_second": 13.392,
      "eval_steps_per_second": 1.7,
      "eval_weighted_fpr": 0.08923884514435695,
      "eval_weighted_sensitivity": 0.7728285077951003,
      "eval_weighted_specificity": 0.8457697048039892,
      "step": 448
    },
    {
      "epoch": 2.232142857142857,
      "grad_norm": 10.557605743408203,
      "learning_rate": 4.631696428571429e-05,
      "loss": 0.8855,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7305122494432071,
      "eval_f1_macro": 0.5313357935186093,
      "eval_f1_micro": 0.730512249443207,
      "eval_f1_weighted": 0.7208063173961943,
      "eval_loss": 0.8126917481422424,
      "eval_macro_fpr": 0.1288106780460812,
      "eval_macro_sensitivity": 0.5599513761278467,
      "eval_macro_specificity": 0.8878517309481255,
      "eval_precision": 0.7320572933103704,
      "eval_precision_macro": 0.5335863697705803,
      "eval_recall": 0.7305122494432071,
      "eval_recall_macro": 0.5599513761278467,
      "eval_runtime": 35.5026,
      "eval_samples_per_second": 12.647,
      "eval_steps_per_second": 1.606,
      "eval_weighted_fpr": 0.10950226244343891,
      "eval_weighted_sensitivity": 0.7305122494432071,
      "eval_weighted_specificity": 0.8208946743492954,
      "step": 672
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.779510022271715,
      "eval_f1_macro": 0.5605264699901076,
      "eval_f1_micro": 0.779510022271715,
      "eval_f1_weighted": 0.756140232373208,
      "eval_loss": 1.0185617208480835,
      "eval_macro_fpr": 0.11840141313918778,
      "eval_macro_sensitivity": 0.565439197792139,
      "eval_macro_specificity": 0.8949862362922303,
      "eval_precision": 0.7503024658142448,
      "eval_precision_macro": 0.5721560130010834,
      "eval_recall": 0.779510022271715,
      "eval_recall_macro": 0.565439197792139,
      "eval_runtime": 42.1679,
      "eval_samples_per_second": 10.648,
      "eval_steps_per_second": 1.352,
      "eval_weighted_fpr": 0.08616187989556136,
      "eval_weighted_sensitivity": 0.779510022271715,
      "eval_weighted_specificity": 0.800434922897206,
      "step": 896
    },
    {
      "epoch": 4.464285714285714,
      "grad_norm": 4.521566867828369,
      "learning_rate": 4.260416666666667e-05,
      "loss": 0.5551,
      "step": 1000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8084632516703786,
      "eval_f1_macro": 0.5892045454545454,
      "eval_f1_micro": 0.8084632516703786,
      "eval_f1_weighted": 0.7866926503340758,
      "eval_loss": 0.7591320276260376,
      "eval_macro_fpr": 0.09875084526504571,
      "eval_macro_sensitivity": 0.5962627065568242,
      "eval_macro_specificity": 0.91148202322346,
      "eval_precision": 0.7673786757699211,
      "eval_precision_macro": 0.583308107932168,
      "eval_recall": 0.8084632516703786,
      "eval_recall_macro": 0.5962627065568242,
      "eval_runtime": 42.7116,
      "eval_samples_per_second": 10.512,
      "eval_steps_per_second": 1.335,
      "eval_weighted_fpr": 0.07319148936170213,
      "eval_weighted_sensitivity": 0.8084632516703786,
      "eval_weighted_specificity": 0.8374648412234613,
      "step": 1120
    }
  ],
  "logging_steps": 500,
  "max_steps": 6720,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "save_steps": 500,
  "total_flos": 2357517389660160.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}