From the Github description:
Deep learning-based tool that allows for various types of new sample generation, as well as sound classification, and searching for similar samples in an existing sample library.
For the comment asking about audio samples, I'm guessing that would be relevant only to the sample generation part of the tool. The rest seems to be more for organizing and finding samples.
There is a Jupyter notebook with examples: https://github.com/maxfrenzel/SampleVAE/blob/master/Examples...
Also curious how the organising and finding part works on generated hybrids.