run-33/checkpoint-1032/trainer_state.json

{
  "best_metric": 0.8166666666666667,
  "best_model_checkpoint": "Bert_tinybert-distilled/run-33/checkpoint-860",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1032,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.8714374899864197,
      "learning_rate": 0.0001496514315519773,
      "loss": 0.4869,
      "step": 172
    },
    {
      "epoch": 1.0,
      "eval_AUC": 0.8238275555555555,
      "eval_BAC": 0.7426666666666666,
      "eval_MCC": 0.48550598838444897,
      "eval_PR_AUC": 0.7856295371225543,
      "eval_accuracy": 0.7426666666666667,
      "eval_loss": 0.44039902091026306,
      "eval_runtime": 6.2687,
      "eval_samples_per_second": 239.283,
      "eval_steps_per_second": 7.498,
      "step": 172
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.5228420495986938,
      "learning_rate": 0.00012827265561598055,
      "loss": 0.4538,
      "step": 344
    },
    {
      "epoch": 2.0,
      "eval_AUC": 0.8575324444444444,
      "eval_BAC": 0.7833333333333333,
      "eval_MCC": 0.5945533134180191,
      "eval_PR_AUC": 0.8074672778987287,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.4205223023891449,
      "eval_runtime": 6.2791,
      "eval_samples_per_second": 238.889,
      "eval_steps_per_second": 7.485,
      "step": 344
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.2492823600769043,
      "learning_rate": 0.00010689387967998379,
      "loss": 0.4415,
      "step": 516
    },
    {
      "epoch": 3.0,
      "eval_AUC": 0.8663128888888889,
      "eval_BAC": 0.8046666666666666,
      "eval_MCC": 0.6114945050025467,
      "eval_PR_AUC": 0.8316399526656422,
      "eval_accuracy": 0.8046666666666666,
      "eval_loss": 0.40513432025909424,
      "eval_runtime": 6.2792,
      "eval_samples_per_second": 238.886,
      "eval_steps_per_second": 7.485,
      "step": 516
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.7223961353302,
      "learning_rate": 8.551510374398703e-05,
      "loss": 0.4324,
      "step": 688
    },
    {
      "epoch": 4.0,
      "eval_AUC": 0.8720817777777777,
      "eval_BAC": 0.8026666666666666,
      "eval_MCC": 0.6129675036984655,
      "eval_PR_AUC": 0.8433174330584201,
      "eval_accuracy": 0.8026666666666666,
      "eval_loss": 0.40090322494506836,
      "eval_runtime": 6.223,
      "eval_samples_per_second": 241.04,
      "eval_steps_per_second": 7.553,
      "step": 688
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.3804311752319336,
      "learning_rate": 6.413632780799028e-05,
      "loss": 0.4243,
      "step": 860
    },
    {
      "epoch": 5.0,
      "eval_AUC": 0.8757084444444445,
      "eval_BAC": 0.8166666666666667,
      "eval_MCC": 0.6384760414061431,
      "eval_PR_AUC": 0.841910429701927,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.3949587345123291,
      "eval_runtime": 6.2206,
      "eval_samples_per_second": 241.134,
      "eval_steps_per_second": 7.556,
      "step": 860
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.10382080078125,
      "learning_rate": 4.2757551871993515e-05,
      "loss": 0.4214,
      "step": 1032
    },
    {
      "epoch": 6.0,
      "eval_AUC": 0.8779199999999998,
      "eval_BAC": 0.8133333333333334,
      "eval_MCC": 0.6351258565198358,
      "eval_PR_AUC": 0.8459677455056573,
      "eval_accuracy": 0.8133333333333334,
      "eval_loss": 0.39859509468078613,
      "eval_runtime": 6.2457,
      "eval_samples_per_second": 240.165,
      "eval_steps_per_second": 7.525,
      "step": 1032
    }
  ],
  "logging_steps": 500,
  "max_steps": 1376,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "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": 2702266380000.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.7463444212571515,
    "learning_rate": 0.00017103020748797406,
    "num_train_epochs": 8,
    "temperature": 2
  }
}