Hands-On Python Deep Learning for the Web

The MNIST dataset of handwritten digits

This dataset was created by a team made up of Yann LeCun, Corinna Cortes, and Christopher J.C. Burges. It is a large collection of images of handwritten digits, containing 60,000 training samples and 10,000 testing samples. The dataset is publicly available for download at http://yann.lecun.com/exdb/mnist/ where it is present in the form of four .gz compressed files.

The four files are as follows:

• train-images-idx3-ubyte.gz: The training set images. These images will be used to train the neural network classifier.
• train-labels-idx1-ubyte.gz: The training set labels. Every image in the training set will have a label associated with it, which is the corresponding digit visible in that image.
• t10k-images-idx3-ubyte.gz ...