YuE: Open Full-Song Generation Foundation Model

And here we are, we finally have a Suno v2 model available for free. Not even music generation has a moat now. I know Suno reached V4 which is incredibly great, but having this open source feels incredible.

I would love to see if this could be quantized to run on lower-end hardware (like 16GB VRAM) with decent output. Instead of using TTS you could generate a short 15s audio jingle when a program runs or a command ends.