Keras Installation and API

In the previous chapter, we discussed the basic principles of neural networks and provided a few examples of nets that are able to recognize MNIST handwritten numbers.

This chapter explains how to install Keras, Theano, and TensorFlow. Step by step, we will look at how to get the environment working and move from intuition to working nets in very little time. Then we will discuss how to install on a dockerized infrastructure based on containers, and in the cloud with Google GCP, Amazon AWS, and Microsoft Azure. In addition to that, we will present an overview of Keras APIs, and some commonly useful operations such as loading and saving neural networks' architectures and weights, early stopping, history saving, checkpointing, and interactions with TensorBoard and Quiver. Let us start.

By the end of this chapter, we will have covered the following topics:

• Installing and configuring Keras
• Keras architecture