run-10/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.8256880733944955,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-10/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,
      "grad_norm": 14.234737396240234,
      "learning_rate": 7.828321466767788e-05,
      "loss": 1.9973,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8165137614678899,
      "eval_loss": 1.6924320459365845,
      "eval_runtime": 2.8153,
      "eval_samples_per_second": 309.737,
      "eval_steps_per_second": 2.486,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 18.908693313598633,
      "learning_rate": 1.482636641433293e-07,
      "loss": 1.0258,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8256880733944955,
      "eval_loss": 1.6462146043777466,
      "eval_runtime": 2.8215,
      "eval_samples_per_second": 309.061,
      "eval_steps_per_second": 2.481,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 1054,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 16141525436100.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.015302977727103428,
    "learning_rate": 0.00015626990200706908,
    "num_train_epochs": 2,
    "temperature": 5
  }
}