GLiMORiA
A successor of my previous SDXL LoRA that had the same name and dataset. Flux has allowed the images that are generated to more closely match the dataset, which makes me quite happy.
The idea just came to me one day a while ago. The style is 2010s anime, the aesthetics are Victorian, the colors are fluorescent, and the somewhat discordant (but pleasing) mixture of these things is complimented by elements of glitch art.
This model is trained with a dataset allowing for portraits of people of varying appearances, although the clothing will almost always be Victorian-era. This model is also trained on generating multiple types of animals (usually with cute lace collars and/or hats). Further, it produces beautiful cityscapes which are, you guessed it, overwhelmingly Victorian.
(GLi)tch + ani(M)e + Vict(ORiA)n = GLiMORiA
Thank you to @NinjawiththeBrZ for his patronage!
๐๐ฐ๐๐๐๐๐๐๐๐ด๐๐๐๐๐๐
Usage guide
Optimal prompting keywords
We recommend using these trigger words when prompting GLiMORiA to get the best results:
- fluorescent colors such as "electric blue" or "hot pink"
- "cityscape" will generate a nice glitchy Victorian cityscape
- most visual descriptions of a person that one can think of
- most species and variations of animals
- just the LoRA tag + overall prompt weight at 100% = fun
Hey Indy! I'm planning on making a LoRA in where to make the look I'm trying to get I will have to use Flux LoRA to create the dataset, Can I use this LoRA in the dataset? I will credit both this LoRA and you ๐
Comments