13EXPERIMENTS WITH KERAS AND MNIST
In the last chapter, we covered the essential components and functionality of a CNN. In this chapter, we’ll work with our test model from Chapter 12. We’ll first learn how to implement and train it in Keras. After that, we’ll conduct a set of experiments that will build our intuition for how different architectures and learning parameter choices affect the model.
From there, we’ll move beyond classification of simple input images and expand the network by converting it into a fully convolutional model capable of processing arbitrary inputs and locating digits wherever they occur in the input.
After fully convolutional ...
Get Practical Deep Learning now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
