Compared to version 1.0, version 2.0 has increased the dataset resolution to 2k and made some adjustments to tagsThe probability of watermark appearing decreasesBetter texturePlease use ADetailer2.0版本较1.0版本，数据集分辨率扩大到2k，tags也做了一定调整水印出现的概率变少质感更好使用ADetailerI trained this Lora using FALLEN LADY's character atlas, which can change the overall art style towards 3D direction, with more images than the previous atlas and better character texture. The images in the dataset are all monochromatic backgrounds, so this Lora is relatively weak in terms of background我用FALLEN LADY的角色图集训练这个lora，可以把整体的画风往3D方向改变，比之前的图集数量多一些，人物质感更好，数据集的图片都是单色背景，所以这个Lora在背景方面比较弱

Files with SHA256: a1c1025cded70e76d1e35d844353fb1b77553e2f24a01cbd21b48846a536a11e