Have you ever wanted theoretically infinite variations of the controversial rapper Yolandi Vi$$er? Does the sight of her jowly scowling visage and or midriff capture you? Have you ever wanted to be sold into slavery? Well, here is the quick and dirty Yolandi Vi$$er LoRA for your pleasure. Trained on 95 images of the quite possibly satanic South African hip hop artist. Dataset was scraped from the Brave search engine and combed for duplicates and false matches. Caption strategy was to use BLIP captions with 8 beams maxing out at roughly 30 characters for concise descriptions.
EDIT: I'd like to stress that these claims are allegations.
Description
FAQ
Comments (4)
It looks like you got really good results. Any tips you can give for your lora creation process? Is this just ~1600 steps constant learning?
I never expected Child Slavery from them...
Cant download the file everytime error