JSFakes (Dr. Tristan Behrens).

This is a tokenized version of the JS-Fakes dataset by Omar Peracha. The original dataset can be found here: js-fakes.git

The representation is four tracks with four bars per track.

Purpose.

This dataset is a good starting point for Music Generation. You could train GPT-2 on the samples to compose music.

Contact.

Find me on LinkedIn and say hello.

If you find and issue or have a feature request, please contact me.

Please be so kind and like this dataset if you find it useful.