Clothing LoRA of Grea's "Light" version from Granblue Fantasy.Hi-Res fix HIGHLY recommended.This LoRA can reproduce Grea without <a target="_blank" rel="ugc" href="https://civit.ai/models/608737">my Grea LoRA</a> (there isn't that much difference - my LoRA helps a bit with hair, horn shape and tail) by using <code>dragon girl, dragon horns, dragon wings, dragon tail,</code>(Use them in negative should these things appear unwanted)<h3 id="main-trigger:-gl_dress-ot5y4ktni">Main trigger: gl_dress</h3>See info sidebar for main tags per version (v0.2b has more than v0.1a)Other useful tags:P: <code>tail ornament</code> for character with non-fluff tail (unstable)P/N: <code>hair ornament</code> for the wing-like hair/horn ornaments if they don't show (might need a few tries to get right. Alternatively put in negative. The tiara should work without these)N: <code>clothes pull, pulling own clothes</code> if wrist cloth is pulled instead of attached to bridal gauntletsN: <code>thigh strap</code> can appear from using laceIssues based on my testing:<ul><li>The cloth hanging from wrists is not stable (bit more stable with <code>hand up</code>; more stable with <code>spread arms</code>)</li><li>The cross badge thing on the belly is kinda unstable (mostly outside of upper body/cowboy shots) (worse in v0.1a)</li><li>Pelvic curtain part is biased towards keeping legs together/crossed in upper body shots if that matters (worse in v0.1a)There is little playroom to create a thigh gap using some of these tags and maybe with higher weight, v0.2b doesn't seem to have that issue on certain concepts:<ul><li>P: <code>thigh gap, legs apart</code></li><li>N: <code>legs together, knees together feet apart, crossed legs</code></li></ul></li></ul>Any of these tags could help although they shouldn't be needed: <code>toeless footwear, high heels, jewelry, sideless outfit, backless outfit</code>

Grea Light Outfit | Granblue Fantasy Clothing

Alpha version. First time making Clothing LoRA and a LoRA with a very limited amount of images with quite a few inconsistencies across images.<s>I'll also probably look into training some type of LyCORIS when i get the time </s>That exists now but its... pretty bad. its a bit more stable on some things but overall its much worse than this version so i decided to leave it unpublished. I will try to train another version but pruning only irrelevant tags, this will result in a model that requires a lot of tokens in prompt but might help i hope.

Files with SHA256: 234ed3678a789471238bd232acf98fe16545d390c8deb52f5b1e5bd7e19b8258