请勿将本模型及生成的图片用于任何商业用途

采用了新中式婚服大类的27张图片构成的训练集。同时，为了保证衣服花纹的完整性，对部分full_body图进行切分，形成27张半身图。

为了不对图片中包含的真人形象造成侵权，在处理原始素材时，对所有人脸部分进行了纯色遮盖。所有范例图片均经过了ADetailer插件的面部重绘处理。

(示例图中第七张以后使用了openpose, 因此仅凭提示词无法复现)

推荐的同时使用“JIAYI55”, “JIAYIQS25”。或单独使用“JIAYI55”。绘制全身时，positive部分要加入“(full_body)"，negative部分要加入“(out_of_frame)”," (head out of frame)", "(???? legs)"，否则可能只会得到半身图片。

推荐权重0.6/0.65/0.7

很久没有训练过Lora了，正好看到有比较优质的婚服类的图片，小小练一下手。

Please do not use this model and the generated images for any commercial purposes.

A training set consisting of 27 images from the broad category of new Chinese style wedding dresses is used. Meanwhile, in order to ensure the completeness of the dress pattern, part of the full_body images are sliced to form 27 half-body images.

In order not to infringe on the images of the real people included in the images, all human face parts were masked with solid colors when processing the original source images. Therefore, all example images in gallery have been processed for facial redrawing with the ADetailer plugin.

(Openpose is used after the seventh picture in the example, so it cannot be reproduced by the prompt word alone)

It is recommended to use "JIAYI55" and "JIAYIQS25" at the same time. Or use "JIAYI55" alone. When drawing the whole body, add "(full_body)" to the positive part and "(out_of_frame)"," (head out of frame)", "(???? legs)" to the negative part. , otherwise you may only get a half-body image.

Recommendation weight 0.6/0.65/0.7

It's been a long time since I trained Lora last time. Just happened to see that there are good quality images in the Chinese style wedding clothes category.