This SDXl LoRA was inspired by the model coco-Illustrious-NoobXL-Style by @COCO_OIOI01, because it has 30 sampling steps, I decided to make a LoRA from the best images made by this model
LoRA weights:
-2 to -0.1: plain pictures, basic details, not dynamic views, little to no highlights, plain eyes and simple background
0: a few hair highlights, more details, more unique outfits
0.1 until 0.7: BEST RANGE: complex details and colors, dynamic views, most unique outfits, better lighting
1+: blurry faces, no distinct lines, detailed but unrecognisable, bad anatomy
I suggest to use a Detail Enhancer about lighting (I have created great results with "Cinematic Lighting" and "Perfect Lighting"), but most prompts about overall quality makes the line blurry so you might require a line slider or "sharp_style"
Most of the images trained were by @psi, I am just the one who made it into a LoRA. (So I suppose I could call this Psi LoRA.)
Trained by PHΘΞN1X
