For business inquires, commercial licensing, custom models, and consultation contact me under yamer@rundiffusion.com
If you have enjoyed any of my models please consider donating here: Thank you!
This LoRA is best used with my other SDXL checkpoints: Unstable Diffusers, Realistic V5 and Anime stage รNA as an example, this model reinforces the style with my models, or you can use it with other checkpoints to bring my style along or push up the quality of images.
The original creator of the LoRA is n15g, without him this model would not exist. ๐ธ
The LoRA works best with abstract work and singular character images (with only 1 main individual), it can be used to create colorful chaotic art in weird aesthetic ways, use it to generate colorful images.
It is trained using 427 images (Version 3) generated using multiple checkpoints in various styles that represent my style and give your image a fantastic unique look that no other LoRA can match.
All the images the model was trained on are 576x1024 but that doesn't stop the model from being used on other images sizes.
LoRA V2 vs V3 vs no LoRA
Full image resolution: https://civitai.com/posts/1115319
Resolution and steps: 576x1024, 1024x1024
Steps: 30 to 150
CFG: 5 to 25
Enjoy and have fun!