{
  "best_metric": 0.0399848073720932,
  "best_model_checkpoint": "BioM-ELECTRA-Large-SQuAD2-XLB/checkpoint-528",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 528,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.3008430600166321,
      "eval_runtime": 32.8152,
      "eval_samples_per_second": 42.785,
      "eval_steps_per_second": 5.363,
      "step": 176
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.0839056447148323,
      "eval_runtime": 32.8639,
      "eval_samples_per_second": 42.722,
      "eval_steps_per_second": 5.355,
      "step": 352
    },
    {
      "epoch": 2.840909090909091,
      "grad_norm": 15.312241554260254,
      "learning_rate": 1.0606060606060608e-06,
      "loss": 0.4144,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.0399848073720932,
      "eval_runtime": 32.8671,
      "eval_samples_per_second": 42.717,
      "eval_steps_per_second": 5.355,
      "step": 528
    }
  ],
  "logging_steps": 500,
  "max_steps": 528,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 2933785388132352.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}