UPDATE 24/07/2025 - With the ban on UK users accessing Civitai about to come in I would like to say this remains my favourite model, at times working on this model stressed me out, I would spend hours a day working on it and it got to the point where I was creating custom outfits in Koikatsu just to try and get details pinned down and I think by the end of it I may have nearly hit triple digits of failed test versions that went into the bin before the culmination of all that trial and effort led to the creation of "v7.0 but actually v3.0" which I consider to be the "Perfect LoRA" as far as training this character could go with the material available to train itI would like to thank everyone who has supported this LoRA. While you can donate Buzz to the model if you wish I will obviously not be able to use that Buzz (I would actually refund peoples Buzz back to them) so feel free to support Civitai itself by using whatever Buzz you would have donated to me to generate images withThank you and God Save The KingEureka from Eureka Seven trained for Illustrious(TAGS)Character - Eur, purple eyes, aqua hair, short hairOptional - hairclipOptional - chokerOutfit - blue dress, white dress, uniform, long sleeves, thigh strap, knee boots, white footwearOutfit - black dress, grey dress, uniform, long sleeves, holster, thigh boots, grey footwearOutfit - headgear, white bodysuitOutfit - strapless bra, two-tone bra, two-tone pantiesOptional - insect wings

Eureka [Illustrious] - Eureka Seven (4 Outfits)

What happened to the 3, 4, 5 and 6 versions? Well you see me being the complete idiot that I am I forgot I actually separated the Pony and Illustrious versions to prevent something like this happening and instead did the very thing I was trying to prevent by naming it v7 because there's 6 of the Pony versionsThe /// in the triggers is just a line/prompt break to separate the outfits and wings, remove the /// and delete the bits of the prompt you don't want like you would do when filling in a multi choice form<ul><li>Upscaled all images to at minimum 1024x</li><li>Touched up images in photoshop to remove bits here and there which could cause issues</li><li>She seems more stable than any previous version when doing test images</li><li>The bodysuit and wings are somewhat a hit and a miss but seem better slightly</li><li>She can wear other outfits but as with all LoRA's other outfits will be heavily influenced by the dataset as you might see when you try to generate other dresses but she does appear to be a bit more flexible in generating other outfits</li></ul>

Files with SHA256: ba38364ea68a73c6128c92701e238164d4af8b23940cfd197ef51a1193d44de5