In the spirit of engaging with Mastodon over Twitter, here's our latest work on music audio understanding...
Looking for an audio embedding benchmark for music? Working on a music ML task and need a good baseline? Supporting code for our #ismir2022 paper is now available, including the MULE model:
https://github.com/PandoraMedia/music-audio-representations
Check out the paper here:
https://arxiv.org/abs/2210.03799
W/ @krzn , S. Oramas, F. Gouyon and A.F. Ehmann
Results and Models for Learning Audio Representations of Music Content - GitHub - PandoraMedia/music-audio-representations: Results and Models for Learning Audio Representations of Music Content
It will be part of a longer tutorial with Rostamzadeh and Anna Huang that will include visual and music art generation.
The full tutorial will be live at #NeurIPS on Dec 5 (virtual) https://nips.cc/virtual/2022/events/tutorial