Jellinator v3
Lora for making strange gelatinous creations
A Flux-trained version of the now depreciated Jellinator V.2, which was an SD-XL trained version of the original Jellator Lora for SD1.5. This version is trained on an expanded dataset that includes pictures from all three models.
Able to create transparent, vibrantly colored gelatin versions of most objects and people. Works best for table top pieces. I have found that forcing the AI to render gelatinous transparent textures really makes it run hard, so upscaling these takes longer sometimes.
Usage guide
Works best if preceding two words before tag are “x-shaped”, “x” being the object you are trying to depict. Responds well to modifiers like “transparent, translucent, ethereal, multicolor, opaque, opalescent, luminous, gelatinous”
No comments yet.
No comments yet