deepjazz

Note: deepjazz is no longer being actively developed. It may be refactored at some point in the future. Goodbye and thank you for your interest

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. Check out deepjazz's music on SoundCloud!

Dependencies

• Keras
• Theano ("bleeding-edge" version on GitHub)
• music21

Instructions

Run on CPU with command:

python generator.py [# of epochs]

Run on GPU with command:

THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python generator.py [# of epochs]

Note: running Keras/Theano on GPU is formally supported for only NVIDIA cards (CUDA backend).

Note: preprocess.py must be modified to work with other MIDI files (the relevant "melody" MIDI part needs to be selected). The ability to handle this natively is a planned feature.

Author

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

Citations

This project develops a lot of preprocessing code (with permission) from Evan Chow's jazzml. Thank you Evan! Public examples from the Keras documentation were also referenced.

Code License, Media Copyright

Code is licensed under the Apache License 2.0
Images and other media are copyrighted (Ji-Sung Kim)