lucidrains/musiclm-pytorch

How to add embeddings or adding support for melody embeddings

raghavpatnecha opened this issue · 1 comments

I am able to train and run the modal with text. It would be great if we can add melody embeddings as mentioned on the examples page of google research

Where'd you get the Dataset from? I think this is the one https://research.google.com/audioset/download.html they used but I'm not sure I should use the CSV or the other format.