Using Keras & Theano for deep learning driven jazz generation

I built deepjazz in 36 hours at a hackathon. It uses Keras & Theano, two deep learning libraries, to generate jazz music. Specifically, it builds a two-layer LSTM, learning from the given MIDI file. It uses deep learning, the AI tech that powers Google's AlphaGo and IBM's Watson, to make music -- something that's considered as deeply human.

deepjazz has been featured in The Guardian, Aeon Magazine, Inverse, Data Skeptic, the front page of HackerNews, and GitHub's trending showcase (1200+ stars). It has led to the most popular "AI" artist on SoundCloud with 172,000+ listens. Currently, deepjazz is being used as reference material for the course "Interactive Intelligent Devices" at the University of Perugia.

Want to listen?


Ji-Sung Kim
Princeton University, Department of Computer Science
hello (at)


This project develops a lot of preprocessing code (with permission) from Evan Chow's jazzml.

As seen in

The Guardian, HackerNews, Inverse, Data Skeptic, and GitHub's Trending Showcase