20% is okay rest is bad..no vae
#4
by
VeteranXt
- opened
what are you using for inference?
I've tried both diffusers and default A1111 web ui settings and the output is always comparable to this from the hosted inference API on the front page:
I never got anything like what you posted with the model but sometimes other models gave me results like that when something was not working properly, like xformers not working as it should.. please check your settings, looks like something is not working on your end. vae is linked in the description
proxima
changed discussion status to
closed
Fixed it. I just needed to restart.