This project aims to generate music in the style of the Nintendo Entertainment System (NES) using deep learning. The corpus used for training data is available at https://github.com/chrisdonahue/nesmdb.
Tensorflow is used for this implementation. Improvements are in the works.