WebAssembly: The Definitive Guide

Chapter 9. Applied WebAssembly: TensorFlow.js

Now, the world don’t move to the beat of just one drum, What might be right for you, May not be right for some.

Theme to Diff’rent Strokes

This is the first of our “Applied WebAssembly” chapters, where I highlight potential use cases for the technology. As you will see over the course of the book, there is no single use. Instead, the designers have crafted a platform with increasing reach into just about all aspects of the software development industry. So follow along. I will not be teaching new features per se. Instead, I hope to help shape your understanding of how our industry is changing rapidly and how WebAssembly will assist.

To begin, I want you to stop for a moment and think about programming languages and machine learning. What is the first programming language that comes to mind? There is probably a good chance your answer was not JavaScript. Why would it be? Machine learning is an incredibly performance-oriented activity with monumental computational workloads these days.

Languages such as Python are much more strongly associated with machine learning than JavaScript is. If we are being honest, that is a bit of a stretch too. Python is a horrible numerics language on its own. It strikes a nice balance, however, of readability, flexible programming style (functional, object-oriented, procedural), and wide-ranging coverage of algorithms, visualization, and data-wrangling features. If you can make it run faster, then ...