run-44/checkpoint-860/trainer_state.json

{
  "best_metric": 0.814,
  "best_model_checkpoint": "Bert_tinybert-distilled/run-44/checkpoint-860",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 860,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.0291366577148438,
      "learning_rate": 0.0002345811638383558,
      "loss": 0.5634,
      "step": 172
    },
    {
      "epoch": 1.0,
      "eval_AUC": 0.8257493333333332,
      "eval_BAC": 0.7493333333333334,
      "eval_MCC": 0.4988440649884173,
      "eval_PR_AUC": 0.7792588674864735,
      "eval_accuracy": 0.7493333333333333,
      "eval_loss": 0.5015578866004944,
      "eval_runtime": 6.2231,
      "eval_samples_per_second": 241.036,
      "eval_steps_per_second": 7.552,
      "step": 172
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.020493507385254,
      "learning_rate": 0.00020525851835856133,
      "loss": 0.5193,
      "step": 344
    },
    {
      "epoch": 2.0,
      "eval_AUC": 0.8491928888888889,
      "eval_BAC": 0.782,
      "eval_MCC": 0.5912326479111619,
      "eval_PR_AUC": 0.8017644489241138,
      "eval_accuracy": 0.782,
      "eval_loss": 0.4762227237224579,
      "eval_runtime": 6.2111,
      "eval_samples_per_second": 241.504,
      "eval_steps_per_second": 7.567,
      "step": 344
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.424774169921875,
      "learning_rate": 0.00017593587287876683,
      "loss": 0.4966,
      "step": 516
    },
    {
      "epoch": 3.0,
      "eval_AUC": 0.8737475555555555,
      "eval_BAC": 0.8033333333333333,
      "eval_MCC": 0.6091019874205584,
      "eval_PR_AUC": 0.8433535637379812,
      "eval_accuracy": 0.8033333333333333,
      "eval_loss": 0.4474146366119385,
      "eval_runtime": 6.1822,
      "eval_samples_per_second": 242.634,
      "eval_steps_per_second": 7.603,
      "step": 516
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.390772342681885,
      "learning_rate": 0.00014661322739897237,
      "loss": 0.482,
      "step": 688
    },
    {
      "epoch": 4.0,
      "eval_AUC": 0.8785333333333334,
      "eval_BAC": 0.8133333333333334,
      "eval_MCC": 0.631647403064903,
      "eval_PR_AUC": 0.8486385119641844,
      "eval_accuracy": 0.8133333333333334,
      "eval_loss": 0.43625757098197937,
      "eval_runtime": 6.2122,
      "eval_samples_per_second": 241.458,
      "eval_steps_per_second": 7.566,
      "step": 688
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.0998780727386475,
      "learning_rate": 0.0001172905819191779,
      "loss": 0.4692,
      "step": 860
    },
    {
      "epoch": 5.0,
      "eval_AUC": 0.8852177777777779,
      "eval_BAC": 0.8140000000000001,
      "eval_MCC": 0.6360574434142463,
      "eval_PR_AUC": 0.8613878408756617,
      "eval_accuracy": 0.814,
      "eval_loss": 0.4329131245613098,
      "eval_runtime": 6.2417,
      "eval_samples_per_second": 240.318,
      "eval_steps_per_second": 7.53,
      "step": 860
    }
  ],
  "logging_steps": 500,
  "max_steps": 1548,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2251888650000.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8802632296149603,
    "learning_rate": 0.00026390380931815027,
    "num_train_epochs": 9,
    "temperature": 3
  }
}