This LoRA is a helper for generating kabedon from the receiver's POV. A few of these are already out there and most checkpoints already support it, but I figured having a LoRA to help reduce tagging and increase concept accuracy would help.
This LoRA can generate kabedon with one or two arms. If you want to one arm version, add tag that have the subject using their one of their arms or some other tag to position the arm.