VRChat Mafuyu

基于animagineXLV3_v30训练，生成Booth上售卖的3D模型Mafuyu图像，新人尝试作品。

推荐权重：推荐1。

使用时tag推荐带上Animation style或者用 (推荐)SDXL Style样式选择器的Anime (动漫)

数据集以1152*1152分辨率进行训练，有33张。

触发词：mafuyu，如果想要默认衣服再加上maid，如果想要另一种衣服就把maid去掉再把mafuyu改成mafuyu white(也许是白色版本数据集较少，还原较差)

关于模型文件这么大真的很抱歉XD，我正在尝试如何在保持质量的情况降低文件大小。

Based on the training of animagineXLV3_v30, a 3D model of Mafuyu is generated for sale on Booth. This is a beginner's attempt at creating the artwork.

Recommendation Weight: 1

When using, tag with Animation style or use the Anime (动漫) selector with (推荐)SDXL Style.

The dataset is trained at a resolution of 1152x1152 and consists of 33 images.

Trigger word: mafuyu. If you want the default clothing, add maid; if you want a different outfit, remove maid and change mafuyu to mafuyu white (perhaps because there are fewer datasets for the white version, the reconstruction quality is lower).

I apologize for the large size of the model files; I'm trying to find ways to reduce file size while maintaining quality.