search

lucidrains / audiolm-pytorch

Implementation of AudioLM, a SOTA Language Modeling Approach to Audio Generation out of Google Research, in Pytorch
MIT License
2.41k stars 256 forks source link

Encodec support #135

Closed LWprogramming closed 1 year ago

LWprogramming commented 1 year ago

Lets us use pre-trained Encodec as a replacement for SoundStream. Works for 24kHz specifically

Tested that it works using my demo script

Usage:

encodec = EncodecWrapper()
# now insert it anywhere you'd insert soundstream = soundstream
lucidrains commented 1 year ago

clapping-leonardo-dicaprio

this is amazing! thank you! 🙏💯 will definitely review it tomorrow or Saturday

lucidrains commented 1 year ago

merged, thank you again!