run-3/checkpoint-527/trainer_state.json

{
  "best_metric": 0.805045871559633,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-3/checkpoint-527",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 527,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 9.870479583740234,
      "learning_rate": 8.893046706823292e-05,
      "loss": 1.9749,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.805045871559633,
      "eval_loss": 1.55028235912323,
      "eval_runtime": 2.6975,
      "eval_samples_per_second": 323.263,
      "eval_steps_per_second": 2.595,
      "step": 527
    }
  ],
  "logging_steps": 500,
  "max_steps": 4216,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 8057768762220.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.31543852233103564,
    "learning_rate": 0.0001016348195065519,
    "num_train_epochs": 8,
    "temperature": 11
  }
}