run-26/checkpoint-96/trainer_state.json

{
  "best_metric": 0.7622309197651663,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-26/checkpoint-96",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 96,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.8396148681640625,
      "learning_rate": 0.00012478512104383658,
      "loss": 0.5871,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7622309197651663,
      "eval_f1": 0.796652719665272,
      "eval_loss": 0.48980674147605896,
      "eval_precision": 0.695906432748538,
      "eval_recall": 0.9315068493150684,
      "eval_runtime": 25.0247,
      "eval_samples_per_second": 40.84,
      "eval_steps_per_second": 1.279,
      "step": 96
    }
  ],
  "logging_steps": 500,
  "max_steps": 960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 235695197280.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.910412613354621,
    "learning_rate": 0.00013865013449315176,
    "num_train_epochs": 10,
    "temperature": 8
  }
}