In the context of nested cross$-$validation, what is the purpose of the inner loop?

Group of answer choices

To evaluate the model's generalization performance

To train the final model

To calculate performance metrics

To select the best model or hyperparameters