This LoRA takes standard prompts and turns them into colorful cubist pieces. It does well with a range of strengths. It does very well with color and ok with black and white.
Description
Base SDXL model . Start with strength of 0.7
FAQ
Comments (4)
It is insane how good this LORA is- I suspect the original critics of the various abstract movements would be amused by how easily the 'machine' replicates Human creative efforts!
Yep, computers are like "squares, and rectangles... I got this!"
This is fun. It takes some effort to get something interesting out. Seems you need a right checkpoint and a right prompt (preferably a short one, with weighting against style token). Also often it takes over and the result has little to do what you asked for. But hey, it is pretty.
Start with straightforward SD XL if you are using this one. Then you can work up the complexity of the prompt. Play around with the strength of it to get the output you want. The Flux one behaves a bit better with prompt adherence at 1, I found this one to need to be lower than 0.7 though the running unicorn example was generated on Civit with a 1.0. Good luck! 🍀
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.