V2 update contents.重新使用Nai训练, 消除了之前版本经常出现的光影. 应该比之前的版本更好(大概), 用在一些真人模型效果也意外的好.Re-Train using Nai. Eliminated the light and shadow that often appeared in previous versions. It should be better than the previous version (roughly). surprisingly, it perform well in some realistic models.===================================================大家好, 由于本站似乎还没有雪染千纱的lora, 因此起了自己训练lora的想法, 所有的图片基于弹丸论破第二季和第三季的裁剪截图, 也添加了少部分其他图片. 这是我的第一个lora,所以很多地方还不成熟, 比如动漫的色彩和光影似乎对出图产生了较大的影响, 此外动漫中的近景全身素材也较少. 不过鉴于目前没有千纱老师的lora, 发生来献丑了.Hello everyone, since this site does not seem to have Yukizome Chisa's Lora yet, I came up with the idea of training Lora by myself. All the pictures are based on cropped screenshots from the second and third seasons of Danganronpa, and a few other pictures have also been added. This is my first lora, so many aspects are still immature. For example, the color and light and shadow of animation seem to have a greater impact on the drawing. In addition, there are less close-up and full-body materials in animation. However, given that there are currently no Chisa's ’s lora, so I upload this work.建议权重/suggested weight: 0.6-0.9三套服饰(教室服饰, 女仆服饰和机关服饰)可以根据示例图片查看提示词For three suits(school uniform, maid and labcoat), you can reference the prompt words based on picture instances.

[Inch] 雪染 千纱 / yukizome chisa / 雪染 ちさ

重新使用Nai训练, 消除了之前版本经常出现的光影. 应该比之前的版本更好(大概), 用在一些真人模型效果也意外的好. Re-Train using Nai. Eliminated the light and shadow that often appeared in previous versions. It should be better than the previous version (roughly). surprisingly, it perform well in some realistic models.

Files with SHA256: fddde3ed4aea0db7ed242b7cae595ea95960c30e0741a8706ad916fc03f13b64