You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I tried all the possible upscale modes and all of them are changing the noise into a square form, significantly reducing quality or changing the color tone. This is extremely annoying when you are trying to finetune the follow-up sampling step.
Is there any possibility to upscale the latent but kind of loss-less?
My current workaround is to VAE decode it to pixels and encode it back again, but in a bigger context, this is pretty time and memory consuming.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
I tried all the possible upscale modes and all of them are changing the noise into a square form, significantly reducing quality or changing the color tone. This is extremely annoying when you are trying to finetune the follow-up sampling step.
Is there any possibility to upscale the latent but kind of loss-less?
My current workaround is to VAE decode it to pixels and encode it back again, but in a bigger context, this is pretty time and memory consuming.
Beta Was this translation helpful? Give feedback.
All reactions