This is repository contains implementations of Generative Adversarial Networks and Variational Autoencoder, written in Python with Keras (Tensorflow backend). The aim of this project is to compare the performance of this two different approaches generating artificial images.
As dataset images from the CelebA, MNIST and Char74k dataset are used. In the datasets folder there are .pkl files which contain resized and preprocessed images of each data set.
All results, findings and methods are summarized in the Report. More images are given in the results folder.