I trained about a half-dozen inversions using the A1111 dreambooth mode. Most of them were pretty poor, but this Rin model consistently produces amazing outputs.

Sadly, many models and Lora do poorly with darker skinned characters and monster girls. Depending on the model, you may find that you have to add:

• dark_skin

• animal_ears

• (white animal_ear_fluff)

I find that these are less necessary in Anything-based models, but mixes and Lora will often need it.