{
  "best_metric": 0.5450097847358122,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-1532",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 5362,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.27224254608154297,
      "learning_rate": 0.00018865680984213388,
      "loss": 0.0871,
      "step": 766
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5391389432485323,
      "eval_f1": 0.19761499148211242,
      "eval_loss": 0.0780063048005104,
      "eval_precision": 0.7631578947368421,
      "eval_recall": 0.11350293542074363,
      "eval_runtime": 132.2112,
      "eval_samples_per_second": 7.73,
      "eval_steps_per_second": 1.936,
      "step": 766
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6950057744979858,
      "learning_rate": 0.00015721400820177824,
      "loss": 0.0802,
      "step": 1532
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5450097847358122,
      "eval_f1": 0.22885572139303484,
      "eval_loss": 0.07739852368831635,
      "eval_precision": 0.75,
      "eval_recall": 0.1350293542074364,
      "eval_runtime": 132.9676,
      "eval_samples_per_second": 7.686,
      "eval_steps_per_second": 1.925,
      "step": 1532
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.19945429265499115,
      "learning_rate": 0.0001257712065614226,
      "loss": 0.0788,
      "step": 2298
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5283757338551859,
      "eval_f1": 0.15734265734265732,
      "eval_loss": 0.0775846391916275,
      "eval_precision": 0.7377049180327869,
      "eval_recall": 0.08806262230919765,
      "eval_runtime": 133.3593,
      "eval_samples_per_second": 7.664,
      "eval_steps_per_second": 1.92,
      "step": 2298
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.47727036476135254,
      "learning_rate": 9.432840492106694e-05,
      "loss": 0.0776,
      "step": 3064
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5322896281800391,
      "eval_f1": 0.16724738675958187,
      "eval_loss": 0.07602300494909286,
      "eval_precision": 0.7619047619047619,
      "eval_recall": 0.09393346379647749,
      "eval_runtime": 132.0039,
      "eval_samples_per_second": 7.742,
      "eval_steps_per_second": 1.939,
      "step": 3064
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.0770089626312256,
      "learning_rate": 6.28856032807113e-05,
      "loss": 0.077,
      "step": 3830
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5401174168297456,
      "eval_f1": 0.20068027210884357,
      "eval_loss": 0.07548870891332626,
      "eval_precision": 0.7662337662337663,
      "eval_recall": 0.11545988258317025,
      "eval_runtime": 138.6281,
      "eval_samples_per_second": 7.372,
      "eval_steps_per_second": 1.847,
      "step": 3830
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.5817792415618896,
      "learning_rate": 3.144280164035565e-05,
      "loss": 0.076,
      "step": 4596
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.541095890410959,
      "eval_f1": 0.20373514431239387,
      "eval_loss": 0.07543697953224182,
      "eval_precision": 0.7692307692307693,
      "eval_recall": 0.11741682974559686,
      "eval_runtime": 132.1275,
      "eval_samples_per_second": 7.735,
      "eval_steps_per_second": 1.938,
      "step": 4596
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.6458355784416199,
      "learning_rate": 0.0,
      "loss": 0.0754,
      "step": 5362
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5391389432485323,
      "eval_f1": 0.19761499148211242,
      "eval_loss": 0.07505200803279877,
      "eval_precision": 0.7631578947368421,
      "eval_recall": 0.11350293542074363,
      "eval_runtime": 132.5302,
      "eval_samples_per_second": 7.711,
      "eval_steps_per_second": 1.932,
      "step": 5362
    }
  ],
  "logging_steps": 500,
  "max_steps": 5362,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 1649866380960.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.08859371316848064,
    "learning_rate": 0.00022009961148248953,
    "num_train_epochs": 7,
    "temperature": 26
  }
}