rmadnegative v1 and clearvae v1 are needed to replicate thumbnails, this is my personal mix to have clear tv anime coloring and lines
KYUT!
with shadows!
Recipe
-rabbit v2
-lametta v1504
-coppermix
-jitQ_v20
-flexdream v10
-cuteyukimix adorable kemiaomiao
v3 adds:
-amedira clip
-momoHQfinal
-cuteyukimix (only 2 blocks instead)
PD:
all images are made in comfyui, for V3 i changed my settings to Euler + ddim uniform since it looked better for me
Description
rmadnegative v1 and clearvae v1 are needed to replicate thumbnails, this is my personal mix to have clear tv anime coloring and lines
KYUT!
round and round
with shadows!
Recipe
-rabbit v2
-lametta v1504
-coppermix
-jitQ_v20
-flexdream v10
-adorable cuteyukimix kemiaomiao
FAQ
Comments (4)
i finally got the naming right, sorry about that, civ works weirdly with the names i put xD, darker version becomes more rounder and flexible with the cuteyukimix adorable mixing xD
I've been testing it for a while, but seems to be worse than v1. It's less flexible, ignores a lot of tags that v1 handles fine, and seems quality is lower even after postprocessing (adetailer etc).
I think it's cuteyukimix, I had the same issues when I tried that one. I think its text encoder component is messing the mix up.
understandable
it indeed is cuteyukimix, about 3 blocks mixed with the v1, i just wanted rounder faces and better proportions but i did notice the hand quality worsened, been so long and have been playing with sdxl pony anime lately, there's barely any good 1.5 models that i like to mix with my model xD'nt, v2 is just another variant for me, i use many versions of my model depending of what i want, if i want rounder faces i use v2 xD maybe i'll edit the naming to express that, thanks for the comment
Thanks for replying! I understand what you mean, makes sense as a variant.