This project showcases the ability of generative adversarial networks to make summer pictures look like they were taken in winter and vice versa. It consists of three parts:
-
Builds a CycleGAN model for image-to-image translation.
-
Implements a simple command line utility which takes in a path to a model artifact and an image for conversion.
-
Implements the application allowing a user to select, view, and convert images by means of a graphical interface.