For best results, use Euler A

Another in a similar vein to my other genderless models, this one is a lot more versatile and better at anatomy. Trained on 800+ generated images of various configurations of adult bodies / gender traits / body compositions / complexions, etc. It was trained on several categories of tags to mix various traits and styles easily.

Tag Categoties:

Body type - genderless, enby, masculine, feminine. "genderless" and "enby" both prodce different blends / styles of androgynous body type

Style - realistic, semi-realistic, fine art, toon

Specific features (without descriptors) - breasts, pectoral_muscles, flat_chest, tummy, abs, penis, vagina, ass, fellatio, panty_bulge, bbw, hairy, ambiguous_genitalia[neg]. this was mostly experimental, in an attempt to have the model recognize all variations of a given feature with a general tag, and seeing if applying specific descriptors during prompting had the desired effect; so far in testing, it seems like descriptors added later have the desired effect in most cases.

The ambiguous_genitalia[neg] was used on a handful of the source images with undefined / ambiguous genitals. it works sometimes, it'll need to be trained more thoroughly in the next version.

Recommended strength depends heavily on the base model used, 0.2 to 0.8 has worked for me in most cases.