Questions

After this chapter, you will be able to answer the following questions:

1. Why do we want to use more layers?
2. Is a network with more layers automatically slower than a shallower one?
3. How do we know how to stop training the model?
4. Which Keras function can we use to stop the training before the model starts to overfit?
5. How can you normalize the channels?
6. How can you effectively make your dataset bigger and more difficult?
7. Does dropout make your model more robust?
8. If you use data augmentation, would you expect the training to become slower or faster?
9. If you use dropout, would you expect the training to become slower or faster?