training_checkpoints/checkpoint-3215/trainer_state.json

{
  "best_metric": 0.6012751387344262,
  "best_model_checkpoint": "tiny-mistral/checkpoint-1929",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 3215,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.78,
      "grad_norm": 13.069233894348145,
      "learning_rate": 4.7413167444271646e-05,
      "loss": 1.4479,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6498838109992254,
      "eval_f1_macro": 0.45467871431664447,
      "eval_f1_micro": 0.6498838109992254,
      "eval_f1_weighted": 0.6214028764904677,
      "eval_loss": 1.118202805519104,
      "eval_macro_fpr": 0.039027492204575735,
      "eval_macro_sensitivity": 0.4743680961365506,
      "eval_macro_specificity": 0.9731278674075015,
      "eval_precision": 0.6258370043663808,
      "eval_precision_macro": 0.4712458606651466,
      "eval_recall": 0.6498838109992254,
      "eval_recall_macro": 0.4743680961365506,
      "eval_runtime": 52.8829,
      "eval_samples_per_second": 24.412,
      "eval_steps_per_second": 3.063,
      "eval_weighted_fpr": 0.03705525495982948,
      "eval_weighted_sensitivity": 0.6498838109992254,
      "eval_weighted_specificity": 0.9470342001132973,
      "step": 643
    },
    {
      "epoch": 1.56,
      "grad_norm": 42.20966720581055,
      "learning_rate": 4.482115085536548e-05,
      "loss": 0.8133,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6986831913245546,
      "eval_f1_macro": 0.5473679856051097,
      "eval_f1_micro": 0.6986831913245546,
      "eval_f1_weighted": 0.6969580552726153,
      "eval_loss": 1.0854355096817017,
      "eval_macro_fpr": 0.030488809177152586,
      "eval_macro_sensitivity": 0.552765390116517,
      "eval_macro_specificity": 0.9773020581125585,
      "eval_precision": 0.7196764053623382,
      "eval_precision_macro": 0.5876945028566362,
      "eval_recall": 0.6986831913245546,
      "eval_recall_macro": 0.552765390116517,
      "eval_runtime": 93.8123,
      "eval_samples_per_second": 13.762,
      "eval_steps_per_second": 1.727,
      "eval_weighted_fpr": 0.029883997848966736,
      "eval_weighted_sensitivity": 0.6986831913245546,
      "eval_weighted_specificity": 0.9608476803638234,
      "step": 1286
    },
    {
      "epoch": 2.33,
      "grad_norm": 26.4113712310791,
      "learning_rate": 4.223950233281493e-05,
      "loss": 0.5592,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6986831913245546,
      "eval_f1_macro": 0.6012751387344262,
      "eval_f1_micro": 0.6986831913245546,
      "eval_f1_weighted": 0.6997681579317561,
      "eval_loss": 1.6113793849945068,
      "eval_macro_fpr": 0.030358926871943602,
      "eval_macro_sensitivity": 0.5880627120151358,
      "eval_macro_specificity": 0.9773067719065487,
      "eval_precision": 0.7107021565435693,
      "eval_precision_macro": 0.6368062038606529,
      "eval_recall": 0.6986831913245546,
      "eval_recall_macro": 0.5880627120151358,
      "eval_runtime": 96.7771,
      "eval_samples_per_second": 13.34,
      "eval_steps_per_second": 1.674,
      "eval_weighted_fpr": 0.029883997848966736,
      "eval_weighted_sensitivity": 0.6986831913245546,
      "eval_weighted_specificity": 0.9609183872736765,
      "step": 1929
    },
    {
      "epoch": 3.11,
      "grad_norm": 0.03971128165721893,
      "learning_rate": 3.964748574390876e-05,
      "loss": 0.3916,
      "step": 2000
    },
    {
      "epoch": 3.89,
      "grad_norm": 13.533605575561523,
      "learning_rate": 3.705546915500259e-05,
      "loss": 0.2375,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6955848179705655,
      "eval_f1_macro": 0.5698961254173763,
      "eval_f1_micro": 0.6955848179705655,
      "eval_f1_weighted": 0.6923294650516646,
      "eval_loss": 1.7778925895690918,
      "eval_macro_fpr": 0.030968414851007965,
      "eval_macro_sensitivity": 0.5666678921166624,
      "eval_macro_specificity": 0.9768109367511323,
      "eval_precision": 0.7000704251293273,
      "eval_precision_macro": 0.584042243736088,
      "eval_recall": 0.6955848179705655,
      "eval_recall_macro": 0.5666678921166624,
      "eval_runtime": 94.9154,
      "eval_samples_per_second": 13.602,
      "eval_steps_per_second": 1.707,
      "eval_weighted_fpr": 0.030312379483224065,
      "eval_weighted_sensitivity": 0.6955848179705655,
      "eval_weighted_specificity": 0.956579233296421,
      "step": 2572
    },
    {
      "epoch": 4.67,
      "grad_norm": 0.2196340560913086,
      "learning_rate": 3.446345256609643e-05,
      "loss": 0.1586,
      "step": 3000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6948102246320682,
      "eval_f1_macro": 0.5694948674292386,
      "eval_f1_micro": 0.6948102246320682,
      "eval_f1_weighted": 0.691666971595594,
      "eval_loss": 2.175213575363159,
      "eval_macro_fpr": 0.03156565589564158,
      "eval_macro_sensitivity": 0.5798857481261962,
      "eval_macro_specificity": 0.9769938659292313,
      "eval_precision": 0.7011163809357478,
      "eval_precision_macro": 0.5797124163320468,
      "eval_recall": 0.6948102246320682,
      "eval_recall_macro": 0.5798857481261962,
      "eval_runtime": 96.1916,
      "eval_samples_per_second": 13.421,
      "eval_steps_per_second": 1.684,
      "eval_weighted_fpr": 0.03042001235330451,
      "eval_weighted_sensitivity": 0.6948102246320682,
      "eval_weighted_specificity": 0.9600977643064038,
      "step": 3215
    }
  ],
  "logging_steps": 500,
  "max_steps": 9645,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 1.431761087102976e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}