Programming Machine Learning

Building a Neural Network with Keras

Python boasts a few large ML libraries such as PyTorch^[24] and TensorFlow.^[25] They’re complex, and they tend to evolve fast. I won’t use them in this book, or even talk about them. If I did, my code examples and explanations would likely be obsolete by the time you read these pages.

Instead, I’ll use a slimmer and more stable library named Keras.^[26] Keras is a thin layer that sits on top of the “big” libraries and hides them behind a nice, clean programming interface. Keras started out as a tool for prototyping, but it quickly became one of the most popular ways to build neural networks in Python—it was even adopted by TensorFlow as an official interface. With Keras, we can use the heavyweight libraries ...