It's not perfect but as good as it gets with TI.

The embedding requires "woman" in the positive prompt and "man" in the negative prompt.

It can't generate "full body view" with the correct face, it will always mess up. So don't use that in your prompt.

Doesn't work on typical 2D Anime models, only realistic ones.

I might improve the embedding in the future or try to create Lora.