The basic model of the example graph is : MiaoMiao Harem

The art style trained with AI-generated images is generally stable, but when drawing navels and toes on some overly large canvases, multiple navels and multiple toes will appear, so please pay attention to the choice of prompt words. (Not sure whether filling in prompt words such as "multiple toes" and "multiple navels" in the reverse prompt words will affect the probability of this error)

The above are the problems I found in my own testing. I wish you a pleasant experience. You are also welcome to share your works.