In this article, we’ll build a  TensorFlow.js model to recognize handwritten digits with a convolutional neural network.

Introduction

First, we’ll train the classifier by having it “look” at thousands of handwritten digit images and their labels. Then we’ll evaluate the classifier’s accuracy using test data that the model has never seen.

We will make a webpage that uses TensorFlow.js to train a model in the browser. Given a black and white image of a particular size, it will classify which digit appears in the image. The steps involved are:

• Load the data.
• Define the architecture of the model.
• Train the model and monitor its performance as it trains.
• Evaluate the trained model by making some predictions.

What you’ll need

• A recent version of  Chrome or another modern browser that supports ES6 modules.
• A text editor, either running locally on your machine or on the web via something like  Codepen or  Glitch.
• Knowledge of HTML, CSS, JavaScript, and  Chrome DevTools (or your preferred browsers dev tools).
• A high-level conceptual understanding of  Neural Networks. If you need an introduction or refresher, consider watching  this video by 3blue1brown or this  video on Deep Learning in Javascript by Ashi Krishnan.

#ai #machine-learning #artificial-intelligence #tensorflow #javascript