Hello everyone! This is a product of working on retraining my other LoRA, DFunk LoRA. I had redone the entire dataset using IP adaptors. I was aiming to get a specific style of images and this LoRA did not do that for me. However! It does make some really nice images and it has even made me pull certain checkpoints off my "potentially delete" list because I feel like it adds so much life to images. It's kind of graphic novel style, but also a bit punk style. But it's not really either of those for certain situations. I don't really know how to describe it. Review the images to see for yourself how it behaves. Each checkpoint you use it in has a very different outcome. For already illustrative checkpoints it adds more elements and for photo checkpoints it can add more depth or altered colors to the images. More or less it's chaos. Might make you laugh sometimes. But I have only tested it with a handful of prompts, so not sure how it will behave for your use.

I would love to see your generations if you care to share. Even though this LoRA didn't do what I wanted I didn't want to leave it on my hard drive sitting there without sharing it with you all.

Triggering to use: "DFxnk style design of a -insert thing here- on a black background"

It doesn't always do the black solid background correctly, so please do try it without to see how it behaves.

You can also just do DFxnk.

Blurry in the negative helps a ton for me.

The dataset does not include images of any people to help avoid it influencing those beautiful ????? you all make. But the dataset does include lots of animals, specifically cats and gators.

I would love to put more time into Stable Diffusion dataset curating and editing but it's one of my hobbies that doesn't help make an income. If you care to give a follow, favorite, or any other encouragement, it would help me set aside more time for it.