run-8/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.8302752293577982,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-527",
  "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": 10.257322311401367,
      "learning_rate": 0.0005386391253190077,
      "loss": 1.2539,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8302752293577982,
      "eval_loss": 1.5333763360977173,
      "eval_runtime": 2.8188,
      "eval_samples_per_second": 309.352,
      "eval_steps_per_second": 2.483,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": NaN,
      "learning_rate": 0.0004042348653959536,
      "loss": 0.5485,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.801605504587156,
      "eval_loss": 1.7583402395248413,
      "eval_runtime": 2.8148,
      "eval_samples_per_second": 309.788,
      "eval_steps_per_second": 2.487,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 2635,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 16141525436100.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.31906439494819705,
    "learning_rate": 0.0006732989066487597,
    "num_train_epochs": 5,
    "temperature": 15
  }
}