Wow. Anna is a godsend. Hopefully now we get some really good open source music models

First we need good stem splitting

What do you think about the recent SAM audio model by meta? https://ai.meta.com/blog/sam-audio/

Is it realtime?