This repository contains the supporting Python code for my medium blog post which can be found here.
In this support code, I go through several examples of using autoencoders to solve real-world problems. The first example involves denoising images from the fashion MNIST dataset. The second example involves denoising images of scanned documents which contain creases and dark regions to help make them cleaner and more readable. The third example looks at using variational autoencoders to learn a data generating distribution in order to create its own images similar to those within the fashion MNIST dataset.