checkpoint-1688/trainer_state.json

{
  "best_metric": 0.04810093343257904,
  "best_model_checkpoint": "AIvisionGuard/checkpoint-1688",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1688,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3,
      "grad_norm": 5.197747707366943,
      "learning_rate": 2.175824175824176e-06,
      "loss": 0.0429,
      "step": 500
    },
    {
      "epoch": 0.59,
      "grad_norm": 3.6927614212036133,
      "learning_rate": 1.26007326007326e-06,
      "loss": 0.0382,
      "step": 1000
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.3490743637084961,
      "learning_rate": 3.443223443223443e-07,
      "loss": 0.0383,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9820833333333333,
      "eval_loss": 0.04810093343257904,
      "eval_runtime": 119.5303,
      "eval_samples_per_second": 100.393,
      "eval_steps_per_second": 3.137,
      "step": 1688
    }
  ],
  "logging_steps": 500,
  "max_steps": 1688,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 8.369134878375936e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}