Browser-based Models with TensorFlow.js

Overview

Bringing a machine learning model into the real world involves a lot more than just modeling. This Specialization will teach you how to navigate various deployment scenarios and use data more effectively to train your model. In this first course, you’ll train and run machine learning models in any browser using TensorFlow.js. You’ll learn techniques for handling data in the browser, and at the end you’ll build a computer vision project that recognizes and classifies objects from a webcam. This Specialization builds upon our TensorFlow in Practice Specialization. If you are new to TensorFlow, we recommend that you take the TensorFlow in Practice Specialization first. To develop a deeper, foundational understanding of how neural networks work, we recommend that you take the Deep Learning Specialization.

Syllabus

Introduction to TensorFlow.js

Welcome to Browser-based Models with TensorFlow.js, the first course of the TensorFlow for Data and Deployment Specialization. In this first course, we’re going to look at how to train machine learning models in the browser and how to use them to perform inference using JavaScript. This will allow you to use machine learning directly in the browser as well as on backend servers like Node.js. In the first week of the course, we are going to build some basic models using JavaScript and we'll execute them in simple web pages.

Image Classification In the Browser

This week we'll look at Computer Vision problems, including some of the unique considerations when using JavaScript, such as handling thousands of images for training. By the end of this module you will know how to build a site that lets you draw in the browser and recognizes your handwritten digits!

Converting Models to JSON Format

This week we'll see how to take models that have been created with TensorFlow in Python and convert them to JSON format so that they can run in the browser using Javascript. We will start by looking at two models that have already been pre-converted. One of them is going to be a toxicity classifier, which uses NLP to determine if a phrase is toxic in a number of categories; the other one is Mobilenet which can be used to detect content in images. By the end of this module, you will train a model in Python yourself and convert it to JSON format using the tensorflow.js converter.

Transfer Learning with Pre-Trained Models

One final work type that you'll need when creating Machine Learned applications in the browser is to understand how transfer learning works. This week you'll build a complete web site that uses TensorFlow.js, capturing data from the web cam, and re-training mobilenet to recognize Rock, Paper and Scissors gestures.