run-48/checkpoint-96/trainer_state.json

{
  "best_metric": 0.7710371819960861,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-48/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": 3.192322254180908,
      "learning_rate": 0.00010850949486092582,
      "loss": 0.6315,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7710371819960861,
      "eval_f1": 0.8010204081632653,
      "eval_loss": 0.5084298849105835,
      "eval_precision": 0.7082706766917293,
      "eval_recall": 0.9217221135029354,
      "eval_runtime": 27.9086,
      "eval_samples_per_second": 36.62,
      "eval_steps_per_second": 1.147,
      "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": 0.9922126501724915,
    "learning_rate": 0.0001205661054010287,
    "num_train_epochs": 10,
    "temperature": 4
  }
}