---
license: openrail
datasets:
- Gustavosta/Stable-Diffusion-Prompts
---
## Model Description

<!-- Provide a longer summary of what this model is. -->
该模型基于basil mix,dreamlike,ProtoGen等优秀模型微调，融合而来。
用于解决上述模型在绘制亚洲，中国元素内容时，只能绘制丑陋的刻板印象脸的问题。
同时也能改善和减少绘制亚洲，中国元素内容时，得到更接近tags的绘制内容。

# 基于dreamlike微调与AsiaFacemix效果图
![1.png](https://s3.amazonaws.com/moonup/production/uploads/1674043921260-636c3fa9aae2da3c76ba966b.png)
![2.png](https://s3.amazonaws.com/moonup/production/uploads/1674043993544-636c3fa9aae2da3c76ba966b.png)
![37.png](https://s3.amazonaws.com/moonup/production/uploads/1674044034015-636c3fa9aae2da3c76ba966b.png)

# 基于Image to Image效果图
![s1.png](https://s3.amazonaws.com/moonup/production/uploads/1674044372718-636c3fa9aae2da3c76ba966b.png)
![S2.png](https://s3.amazonaws.com/moonup/production/uploads/1674044468671-636c3fa9aae2da3c76ba966b.png)

# 添加国风汉服lora模型
lora-hanfugirl-v1
该模基于真实的汉服照片训练，相对于v1-5，有更细腻美丽的脸部。
![lora-1-05.png](https://s3.amazonaws.com/moonup/production/uploads/1675083455418-636c3fa9aae2da3c76ba966b.png)
![lora-1-0502.jpg](https://s3.amazonaws.com/moonup/production/uploads/1675083514690-636c3fa9aae2da3c76ba966b.jpeg)

lora-hanfugirl-v1-5
该模基于真实的汉服照片训练，相对于v1,对不同的多个模型和不同分辨下的图片兼容性更好。
![lora0501.png](https://s3.amazonaws.com/moonup/production/uploads/1675083569243-636c3fa9aae2da3c76ba966b.png)
![lora0502.png](https://s3.amazonaws.com/moonup/production/uploads/1675083579745-636c3fa9aae2da3c76ba966b.png)