Closed zqevans closed 8 months ago
@zqevans ah, i bumped a minor version though
can't you just pin it?
it had to happen at some point. layernorm biases were shown to make transformer training unstable
I can pin it, it just means the stable-audio-tools library can't start using any features past that point since it's pinned.
Could you at least make them opt-in? That way the default is still the ideal setup, and newer models get that automatically, but old ones can still be used with the right settings turned on.
@zqevans yea, i think that would overly complicate the repo for little gain
why not just anneal out the biases over an extra epoch of training?
@zqevans btw, how are the new audio results looking?
The recent change to remove biases across the model means I can't load my previously-trained x-transformers models.
It would be nice if the previous behavior were the default to remain backwards-compatible, with an option to remove all of the biases