Examples were made with anime model but it should work with any model.
For prompt and settings just drop image you like to PNG info.
Resolution for txt2img: 512x768
ControlNet settings:
Preprocessor: none Model: openpose.
Control mode: Balanced.
Control weight: 0.8 - 1.
Description
To fix some of the issues with background you can use additional depth map or MLSD to get better perspective (as a second ControlNet unit)
FAQ
Comments (3)
do i put this in lora file？
WyvernessRedNov 28, 2023
No, you just drag and drop the pictures of the stick figures into the 'controlnet' section on the stable diffusion web UI
Which file to include it if I use comfyui?