My Links: discord, twitter/ig. Subscribe for FBB images @ https://patreon.com/fitCorderAI
These configs require installing ComfyUI.
Most UI's require fcSuite (those affected include it in the zip.).
"fcFaceFixXLv2" & "fcFaceFixXL&Upres" requires 2 extensions. You will need fcSuite and Face Detailer nodes.
Note: I'm willing to create custom configs for comfyUI on commission. More Here
Description
FAQ
Comments (7)
is this a joke?
Explain your premise here.
Img2Img does not produce great results which is obvious looking at your preview pics. The refiner is not good at producing great results besides the first pass. It has nothing to do with your particular setup, that's just how it is right now. You can go and take a look at all of the img2img upscaler workflows and realize that they produce worse results than the first passes.
@bericbone So that specific workflow is using 0.9XL to compose the image, refined. then send it to a SD1.5 IMG2IMG so you can use 1.5 loras but with a 1024x1024 base image.
SDXL 0.9 is an experimental research model so I'm releasing different workflows for people to experiment with.
Hay,
First of all, thank you for sharing your work, everything worked great until today.
The error is in "Convert Mask to image" node.
HERE IS ERROR CODE:
Failed to validate prompt for output 206:
* MaskToImage 217:
- Return type mismatch between linked nodes: mask, IMAGE != MASK
I found the reason. The ComfyUI Impact Pack has been upgraded, FaceDetailer (pipe) node just needs to be replaced.
Is there a guide or explanation as to what each version does and is for? I'm not clear on it, but maybe I haven't looked closely enough.