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
Here's a workflow that illustrates the issue. I'm wondering if this is a bug, a missing feature, or just something that can't be supported.
Put 500x500px image in ControlNet as a base. Select any processor.
Set A1111 image size to 1000x1000
Generate -- this works fine
Send the result to inpainting
Use ControlNet with the same 500px image to guide inpainting
Generate -- the resulting ControlNet image is squashed, off to one corner, etc. And doesn't apply at all to the inpainted area
When inpainting, I imagined that the lower-res Control net image would be rendered in its entirety, upscaled to match the target, and then only the inpainted portion would apply. However, this isn't what happens.
Is this something that can be supported, or am I thinking about this wrong? Thanks
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
-
Here's a workflow that illustrates the issue. I'm wondering if this is a bug, a missing feature, or just something that can't be supported.
When inpainting, I imagined that the lower-res Control net image would be rendered in its entirety, upscaled to match the target, and then only the inpainted portion would apply. However, this isn't what happens.
Is this something that can be supported, or am I thinking about this wrong? Thanks
Beta Was this translation helpful? Give feedback.
All reactions