very fundamental question here, but how do people get around the inherent biases of danbooru tag based systems like Illustrious?
There are hundreds of artists and hundreds of thousands of images with varying body shapes and types, but the second you add "obese" you compress everything into the same generic body type. Just scrolling through bbwai you can see what I mean.
You can change height, age, character, clothes, etc etc etc, but for me fat distribution seems to just be very same-y and generic regardless of other traits.