| # 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](https://github.com/omarperacha/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](https://www.linkedin.com/in/dr-tristan-behrens-734967a2/) 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. |