Learning TensorFlow.js

Chapter 12. Dicify: Capstone Project

“Everybody has a plan until they get punched in the mouth.”

—Iron Mike Tyson

All of your training has gotten you through a variety of theories and exercises. As of right now, you know enough to come up with a plan to build new and creative uses for machine learning in TensorFlow.js. In this chapter, you’ll develop your capstone project. Rather than learning yet another aspect of machine learning with TensorFlow.js, you’ll start this chapter with a challenge, and you’ll use your existing skills to build a solution that works. From idea to completion, this chapter will guide you through the execution of solving a problem. Whether this is your first book on machine learning or your 10th, this capstone is your time to shine.

We will:

• Research the problem

• Create and augment data

• Train a model that will solve the problem

• Implement the solution in a website

When you finish this chapter, you’ll have applied your skills from beginning to end to solve a fun machine learning project.

A Dicey Challenge

We’ll be using your newfound skill to blur the line between art and science. Engineers have been using machines for impressive visual feats for years. Most notably, the camera obscura technique (as shown in Figure 12-1) allowed mad scientists to trace live scenery with a lens and a mirror.¹

Figure 12-1. Camera obscura

Today, people are making ...