V1.3 - Anima:
- This is the version for Anima, I really liked it.
- Added frame control to prevent the character from moving out of the frame
V1.2:
- Added "ControlNet"
Some people have asked me to share my character workflow.
Why not?
So I refined it and added a randomizer, enjoy!
WARNING! This method does not work well with V-Pred models.
Description
This is the version for Anima, I really liked it.
FAQ
Comments (8)
Would have been nice to see the prompts for the preview images tbh.
I'm trying this out, but where can I get the segm/Face YOLO v3.pt model?
@Slavikson Thanks! The workflow now works without throwing "missing model" errors. But perhaps I'm missing something but I don't "get" how to use this workflow.
The default prompt ("full body, just standing...") properly gives me a full body character. But how is this different from simply using the same prompt on the base model? I was assuming that this workflow is for generating full-body "sprites" of the same character while retaining identity, but when I change Pose and background under 1 - Char Prompt the output is often different.
@wildkrauss This workflow is intended simply for character generation and further improvement of the result with upscaling.
I have experimented with what you describe, in my "Consistency characters" workflow. However, it is for SDXL and has not been updated for a long time. I will try to use some of the techniques from it for Anima.
I don't know if something broke there, but for some reason I was getting a blurry image instead of a normal one, even though it initially generated normally. Lots of confusing processes, I shut down the generation before it finished. It seems to me there's a much simpler way.