15th Lora Release, Aurora! From The Eminence in Shadow - Put everything in Filter for seeing all images!
You can now request a Character's LORA. CLICK HERE
Character's request from my poll + Not so much IA Art on it or didn't see it much, so i created it.
About this Lora:
Trained on 768x Resolution.
I managed to pick 114+ Images at the Best quality possible were used, every images was sorted manually, Majoritary Anime Screencaps, & Some Art, So A Very Good ressemblance of the Official Aurora Anime TV. (And because otherwise Art on it is very Low).
Weight recommended: 0.6-1. - Work perfectly at Weight 1, Best result: 0.7-1
He's Flexibility is Ok (See examples images)
Trigger Words is: AuroraV4 Also i Reducted the number of prompt needed.
AnimeFullFinalPrunedFP16 used, Probably works with Any Anime Model, i used AmbiantAnAnimeMix < Best result, Abyss, YesMix, MidnightMelt.
Again see examples images for a Better Idea, Also click on About This Version at Right Corner for more Info.
Negative Embedding used: EasyNegative, BadPrompt2.
Share you're image in Comments :).
描述:
You can juggle with: Black hair OR Purple hair prompts, pick what you prefer, i find Black Hair more consistent.
If you want to be Flexibile when changing clothes And If Jewelry & Necklace persist try to force in Negative with: ((Jewelry:1.3, Necklace:1.3)) prompt.
Mole under mouth is not 100% Precise, if you want a Fast Fix just drop ur image in PS. (It's very easy put a Black point).
训练词语: AuroraV4
名称: AuroraV4-10.safetensors
大小 (KB): 36979
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
