Using TensorFlow always follows a similar pattern. Set up your input, your graph, and the functions you want to optimize. Train your model, save it, and reuse it for your applications. As it comes with lots of functionalities, there are very few things TensorFlow cannot do. We will also explore some other types of networks in future chapters.

At this point in the book, we have seen the major mode of machine learning: classification. In the cuming two chapters, we will look at techniques used for two specific kinds of data: music and images. Our first goal will be to build a music-genre classifier.