>>106364164
I honestly always have a difficult moment with OpenPose, few times it recognizes the character's position. And few times SDXL pays attention to the openpose map. Literally I'm having better results using as I already mentioned depth canny and lineart at a weight of 0.4