This is a blend of approximately 50 models, some of these models are also blends and I thus can't account for every training set. I have done some finetuning, mostly photography and illustration and my own abstract art.

I would advise against using this model in any kind of commercial application or blend, mostly because I don't know if there's anything in there that wasn't documented in the original blends I incorporated. There's a lot of NSFW training, you might want to negative prompt "?????" or "????" if you're hoping to avoid that for figurative styles.

The latest version has a VAE baked in, sample images were generated with two additional hypernetworks (Georges Seurat and Edward Hopper).

Earlier versions of the model are on Hugging Face.

https://huggingface.co/Echevarian/KompromatDiffusion/tree/main