This is a great improvement building on an existing model! Very much looking forward to the code, and that LoRA!

Stable LM looks like it’s actually CC-BY-SA, which I think would meet the Open Source Definition.

But none of the Stable Diffusion or Stable Audio models are released under such a license.

Yeah, “async is a niche requirement” has long seemed weird to me. Most programs with any degree of interactivity quickly come to a point where they want more than one thing to be happening at once, and they spend some time waiting on external inputs – whether those are network packets or USB events or timers.

Purely synchronous code only suffices for that narrow and most boring category of programs that know all their inputs up-front and don’t accept any other inputs or interactions at all.