Reply to thread

Thanks for sharing those results. Yes, even a local Stable Diffusion with ControlNet has trouble with any sort of action poses (which is why 99% of AI images you see are just of a character standing around).  I had a little luck when I worked with earlier this year, but I had to use 3D models in conjunction with it to get decent poses. [ATTACH=full]145927[/ATTACH][ATTACH]145927[/ATTACH][ATTACH]145928[/ATTACH]