trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 126,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 6.825396825396825e-05,
      "loss": 0.3612,
      "step": 42
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.905688622754491,
      "eval_f1": 0.9257075471698112,
      "eval_loss": 0.23578055202960968,
      "eval_precision": 0.9691358024691358,
      "eval_recall": 0.8860045146726863,
      "eval_roc_auc": 0.9152244795585653,
      "eval_runtime": 1.9731,
      "eval_samples_per_second": 677.118,
      "eval_steps_per_second": 5.575,
      "step": 42
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.492063492063492e-05,
      "loss": 0.1616,
      "step": 84
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9296407185628742,
      "eval_f1": 0.9468325791855204,
      "eval_loss": 0.19503971934318542,
      "eval_precision": 0.9489795918367347,
      "eval_recall": 0.9446952595936795,
      "eval_roc_auc": 0.9223476297968398,
      "eval_runtime": 1.9788,
      "eval_samples_per_second": 675.158,
      "eval_steps_per_second": 5.559,
      "step": 84
    },
    {
      "epoch": 3.0,
      "learning_rate": 1.5873015873015873e-06,
      "loss": 0.074,
      "step": 126
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9258982035928144,
      "eval_f1": 0.9438457175269428,
      "eval_loss": 0.22528469562530518,
      "eval_precision": 0.9486887115165337,
      "eval_recall": 0.9390519187358917,
      "eval_roc_auc": 0.9195259593679458,
      "eval_runtime": 1.9809,
      "eval_samples_per_second": 674.439,
      "eval_steps_per_second": 5.553,
      "step": 126
    },
    {
      "epoch": 3.0,
      "step": 126,
      "total_flos": 2108308886599680.0,
      "train_loss": 0.19893105635567318,
      "train_runtime": 118.1813,
      "train_samples_per_second": 135.605,
      "train_steps_per_second": 1.066
    }
  ],
  "max_steps": 126,
  "num_train_epochs": 3,
  "total_flos": 2108308886599680.0,
  "trial_name": null,
  "trial_params": null
}