Canny Edge instruction #39
barepixels
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
go to CN tab and check Canny
go to img2img tab and use the "Load image(s) to input" button

Click on input tab (top left) you will see your uploaded image
Switch to setting tab
apply your resolution, style, prompt, press generate
This is speed, no refiner

This is quality with refiner

I should have tried with the same seed
Beta Was this translation helpful? Give feedback.
All reactions