I have absolutely no idea what I am doing with people, squishy unpredictable meatbags, let alone LoRAs of them, therefore this comes with no expectations of amazing things.

That being said, it does seem to get at least a bit of the look-n-feel, but if you feel like you could do better, don't let me stop you!

Prompts are trained, they appear to work if you mix them into a scene fairly well.

Same dataset as SDXL, WD14 tagged.