关于更新：在chenkinnoob0.5上训练的lokr，所以文件比较小，不意味更差，灵活性上更好，可以给以前老的模型用关于更新：vpred版适用于vpred的模型，此次更新的版本无需使用启动词。ps:也许你可以通过一些妙妙小工具找到Lora里有什么妙妙tag。这是一个画图非常好看的lora，尤其在搭配了艺术家tag的时候，它产生的作用是让人物的衣褶变得真实并让脸型变得好看，同时给予了好看的打光光线仅使用启动词就是帖子第一张图的效果，如果搭配艺术家tag就是帖子里给了提示词里的图，我给的艺术家风格组合。另外，civitai不让我上传的提示词里有lolita，实际上如果你看到有华丽大裙子的图，那大概是我tag里有Lolita fashion。在我用到的训练集里，我使用了非常多的有lolita服饰的图片，占据了三分之二。因此，这个lora还是一个能提供好看lolita衣服的lora。不过你不加lolita对整个lora起效也是不会有劣化效果的。This is a very aesthetically pleasing Lora for drawing,especially when paired with an artist <a target="_blank" rel="ugc" href="http://tag.It">tag.It</a> makes the characters'clothing creases look realistic and enhances the facial features,while also providing beautiful lighting effects.Using only the activation word will give you the effect of the first image in the post.If you combine it with an artist tag,it will produce the image hinted at in the prompt,which is the artist style combination I provided.Additionally,Civitai doesn't allow me to upload prompts that include"lolita."However,if you see images with elaborate,large skirts,it's likely because my tags include"Lolita fashion."In the training dataset I used,a significant portion,about two-thirds,consists of images featuring Lolita clothing.Therefore,this Lora is also capable of generating visually appealing Lolita outfits.However,even if you don't include"lolita"in the prompt,it won't degrade the overall effectiveness of the Lora.The English introduction here was translated using AI,which may not be accurate.I don't speak English myself,so please forgive me.采样euler 或 euler a 或 euler a karrs推荐强度0.8-1.4虽然是noob其实illustrious也可以用

coserstely

更新的vpred版，无需使用之前的启动词

redshaoji-000031.safetensors

Files with SHA256: 6b7e7a86c1988f48bd08d4142c3ac22f7ab68d6f724ed820ecd46fea45c50f0d