This lab trains and evaluates a handwritten digit classification model using the MNIST dataset. It uses a Graphical Processing Unit (GPU) to speed up training and includes an interactive component that lets you test your model by drawing your own digits right inside this notebook.

Here’s what we’ll do in this lab:

• Train an ML model on a GPU

• Test our model interactively using TensorFlow.js

• Export our model for use in JavaScript and run an interactive web app to test the model interactively, with hand-drawn digits (as depicted in the animated gif above).

Here’s the notebook: mco.fyi/mllab.