Home Creators Posts Import Register Favorites Logout

Downloads

Content

Lots of changes ! Extension class + callbacks / generally updated this to be more in line with my other current operators. - Fixed Seed to be deterministic- Fixed extra model loaded in vram (shouldn't go over ~16 at maximum for musicgen only)- Added unload model command to free up vram but keep server listening- Added support for Audiogen model (although haven't really gotten anything amazing out of it) - Added optional custom Outputfolder (in 'Playback' page)- Moved logic into Extension class - Added Callbacks system- Added interactive playback display mode with view of full audio wave form. 'p' key is hotkey when hovering over, for play/pause- Custom model can be loaded by replacing model files in .cache / Expect a better way to do this soon / message me if you would like this.

Note: For audiogen, you will need to update by hitting Install + Launch musicgen again. and models might redownload for some reason... you can find the model cache here : C:\Users\<username>\.cache\huggingface\hub in case you want to delete old model cache folders.

Comments

Gilles

Did anybody manage to make it run on the 2023 version?

Long

hello i tried to launch the tox but it just keep showing "Traceback (most recent call last): File "C:\Users\ASUS\Desktop\FIle Đã Làm\MusicGen Folder\musicgen_server.py", line 7, in from audiocraft.models import MusicGen, AudioGen ModuleNotFoundError: No module named 'audiocraft' (musicgen_env) C:\Users\ASUS\Desktop\FIle Đã Làm\MusicGen Folder>" How can I fix it ?, thank you