GitLab

This causes some small changes to the uni pc inpainting behavior but it
seems to improve results slightly.

apply_model in model_base now returns the denoised output.

This means that sampling_function now computes things on the denoised
output instead of the model output. This should make things more consistent
across current and future models.

Callback format is: callback(step, x0, x)

On the KSamplers denoise of 1.0 is the same as txt2img but there was a
small difference on UniPC.

The extra step doesn't actually call the unet so it doesn't belong in
the progress bar.