Pony version of this model: https://civarchive.com/models/1001277/saint-ninja-sentai-kage-blue-or-giga-tokusatsu-heroine
After realizing that Illustrious deals with hexagonal patterns better than Pony, I decide to train this LoRA to see if Illustrious can deal with fishnet pattern better. Well, I definitely wouldn't say it's definitely better ^^"
This clothing LoRA is based on the costume design of the recent (as of Dec 2024) iterations of the Kage Blue JAVs from GIGA (specifically THZ-93, SPSB-86, and SPSC-58)
Prompts:
sentaikageblue, shiny clothes, spandex, silver bodysuit, fishnet bodysuit, bodysuit under clothes, aqua dress, shoulder pads, aqua skirt, black belt, aqua glovesFor "from below", "spread legs", etc. images:
aqua pantiesFor "full body" images:
boots, aqua footwear, colored shoe solesHelmet prompts:
helmet, aqua helmet, worn helmethelmet, aqua helmet, no headwear, unworn helmet, holding helmetwhite open face hoodno headwearNotes:
Despite the training data set consisting entirely of aqua color, the AI is able to generate various different colors since I captioned the training data set with "aqua [clothing]". Though if generating blue and green, make sure to negative "aqua [clothing]". Also take note that the shoulder pads, belt, and golden portions of the gloves and boots might also change colors
Illustrious doesn't seem to deal with the "aqua panties" prompt that well in this context of panties over bodysuit (resulting in it generating images without the fishnet bodysuit but rather as thighhighs), so most of the time I either omit it or add it at a very low weight
Didn't test as well as the Pony version with the exact same training data, might want to refer to the Pony version for more notes
Description
FAQ
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.