run-1/checkpoint-75/trainer_state.json

{
  "best_metric": 0.500990099009901,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-1/checkpoint-75",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 75,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.6191820502281189,
      "learning_rate": 5.7462097344528874e-05,
      "loss": 0.3664,
      "step": 75
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.500990099009901,
      "eval_loss": 0.35016587376594543,
      "eval_runtime": 57.4035,
      "eval_samples_per_second": 8.797,
      "eval_steps_per_second": 0.07,
      "step": 75
    }
  ],
  "logging_steps": 500,
  "max_steps": 750,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 777843283140.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.43390314470052016,
    "learning_rate": 6.38467748272543e-05,
    "num_train_epochs": 10,
    "temperature": 19
  }
}