run-1/checkpoint-1069/trainer_state.json

{
  "best_metric": 0.0,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-1/checkpoint-1069",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1069,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.47,
      "grad_norm": 2.2506868839263916,
      "learning_rate": 9.742690012501051e-07,
      "loss": 0.6135,
      "step": 500
    },
    {
      "epoch": 0.94,
      "grad_norm": 3.937321424484253,
      "learning_rate": 1.1814509856987217e-07,
      "loss": 0.5884,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.589656412601471,
      "eval_matthews_correlation": 0.0,
      "eval_runtime": 0.7173,
      "eval_samples_per_second": 1453.971,
      "eval_steps_per_second": 92.006,
      "step": 1069
    }
  ],
  "logging_steps": 500,
  "max_steps": 1069,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 37540018897248.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.8303929039303383e-06,
    "num_train_epochs": 1,
    "per_device_train_batch_size": 8,
    "seed": 30
  }
}