使用方式：

把模型放在embeddings文件夹，并在负向词里填写该模型的名称

How to use:

Put the model in the embeddings folder and fill in the name of the model in the negative word

该模型旨在使SDXL生成的图片具有更好的细节，该模型不具备任何的风格/物体/人物倾向，该模型不会干扰/忽视任何的文本输入，因此该模型能够与任何基于SDXL1.0的Checkpoints，Lora类型的模型共同使用，并尽力提高这些模型生成图片的细节和准确度。

This model is designed to make the pictures generated by SDXL have better details. The model does not have any style/object/character tendencies. The model does not interfere with/ignore any text input, so the model can work with Checkpoints, Lora type models that based on SDXL1.0. This model will try its best to improve the details and accuracy of the pictures generated by these models.

为什么你不需要这个模型：

SDXL即使不使用任何负向词，生成内容的质量也比较高，这与SD1.5已经截然不同。而模型本身并不知道什么是对什么是错，因此，该模型并不一定能永远使得图片质量变得更高，它也有概率使得图片变得更糟糕。而面对一些质量很差的种子，该模型也无能为力。

Why you don't need this model:

Even if SDXL does not use any negative words, the quality of the generated content is relatively high, which is completely different from SD1.5. The model itself does not know what is right and what is wrong. Therefore, this model may not always make the picture quality better, it may also make the picture worse. This model is powerless against some very poor quality seeds.

为什么你需要这个模型：

这个模型确实在一定程度上使得图片变得更清晰，细节更丰富，甚至能修复一定程度上的错误。

Why you need this model:

This model does make the picture clearer and more detailed to a certain extent, and can even repair errors to a certain extent.

该模型由杭州造物云技术有限公司团队开发制作。如需商务合作或有任何疑问，可联系 luping@zaowuyun.com。

The model was developed and created by the team of Hangzhou Virtual Creation Technology Co., Ltd. For business cooperation or any questions, you can contact luping@zaowuyun.com.