run-1/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.8130733944954128,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-1/checkpoint-1054",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1054,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 0.0006875917567735723,
      "loss": 0.9069,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.801605504587156,
      "eval_loss": 1.1160032749176025,
      "eval_runtime": 9.4507,
      "eval_samples_per_second": 92.269,
      "eval_steps_per_second": 0.741,
      "step": 527
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0006111926726876199,
      "loss": 0.4218,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8130733944954128,
      "eval_loss": 1.181414246559143,
      "eval_runtime": 9.1074,
      "eval_samples_per_second": 95.746,
      "eval_steps_per_second": 0.769,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 5270,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 16162868129520.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.5682681228180307,
    "learning_rate": 0.0007639908408595248,
    "num_train_epochs": 10,
    "temperature": 8
  }
}