Because convolutional neural networks work so well with images, to learn more about them, we're going to start with a dataset of images.
The images we're going to work with are from the Food-101 dataset, a collection of 101 different categories of 101,000 (1000 images per category) real-world images of food dishes.
To begin, we're only going to use two of the categories, pizza 🍕 and steak 🥩 and build a binary classifier.