trainer_state.json

{
  "best_metric": 0.03306839242577553,
  "best_model_checkpoint": "BioM-ELECTRA-Large-SQuAD2-XTendons/checkpoint-288",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 288,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.15969941020011902,
      "eval_runtime": 18.3349,
      "eval_samples_per_second": 41.724,
      "eval_steps_per_second": 5.236,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.052657194435596466,
      "eval_runtime": 18.2464,
      "eval_samples_per_second": 41.926,
      "eval_steps_per_second": 5.261,
      "step": 192
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.03306839242577553,
      "eval_runtime": 18.2807,
      "eval_samples_per_second": 41.847,
      "eval_steps_per_second": 5.251,
      "step": 288
    }
  ],
  "logging_steps": 500,
  "max_steps": 288,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1598536910200320.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}