Updated: Jun 28, 2026
concept
Re-up, It's still weird.
SDXL versions.
V1 is tends towards cartoon/CGI type weirdness.
V2 is more illustrative or maybe softer.
V3 I made by mistake with the dataset and captioning for the ZIT version.
Illus version is made with the V1 dataset and tagging.
ZIT made with a pruned version of the V2 dataset captioned instead of tagged.
Tends towards oversaturation with unless CFG is low (<4)