Who think these drawn pics from SD 512 AI? Extremely qualityze v2HQ model have almost "ZWEI" twice resolution than SD1.5. previews specified "child"prompt and tiny life size but this model have all good life size...??? maybe.
High Definition Basemodel
Suited train for full body with as any resolution. Up quality any training(sample is 288x384).
Version History
V1.0: Testing highPixelar2D effection.
V1.1: Prepared exclusive highPixelar LoRA.
V2.0(HQ): Qualityze fill to all resolution(versatile base model).
🌟 How to get true detailed definition feeling on SD1.5 low resolution?(nearly 2.1768 & SDXL feeling)
🌟 Information of NON LCM Fast Method deverlop workflow
https://civarchive.com/articles/3707
Who think these drawn pics from SD 512 AI? Extremely qualityze v2HQ model have almost "ZWEI" twice resolution than SD1.5. previews specified "child"prompt and tiny life size but this model have all good life size...??? maybe.
Original LoRA origin and descriptions are below
儿童绘本插画MOMO(LiblibAI·哩布哩布AI - 中国领先的AI创作平台)
https://www.liblib.art/modelinfo/53793f6bb8a84b609423f1a66b1ea9f1
儿童绘本插画MOMO，卡通动漫动画场景风格LoRA， 欢迎使用返图 感谢大家
加上 illustrations 触发词，简单描述就能很好的出图了。开启高清修复！步数：20左右， 重绘幅度0.6左右 放大算法：R-ESRGAN 4x+ Anime6B
采样方法：Euler a，dpm++2M SDE karras
开启高清修复 放大2倍
权重 0.6-0.95之间使用效果最佳
放大算法：R-ESRGAN 4x+ Anime6B
高分迭代:20 重绘幅度:0.4-0.6
放大倍数 :2 CFG 提示词引导系数:7
正面关键词 例如：((A child,small cute plush ball errings，pigtails，jumper skirt,A charming smile,reading picture-story book)),Parks, lawns, cherry trees, benches
负面关键词：(nsfw:1.5),verybadimagenegative_v1.3,ng_deepnegative_v1_75t,(ugly face:0.8),cross-eyed,sketches,(worst quality:2),(low quality:2),(normal quality:2),lowres,normal quality,((monochrome)),((grayscale)),skin spots,acnes,skin blemishes,bad anatomy,nsfw,