Update MultiControlnet.md

This commit is contained in:
Qianyi Zhao
2024-11-05 03:45:15 -06:00
committed by GitHub
parent e0c09ed53d
commit 99726e02de

View File

@@ -44,8 +44,7 @@ image_2 = pipe(
)
image_2.save("image_6.jpg")
```
|![image_5](https://github.com/user-attachments/assets/19d2abc4-36ae-4163-a8da-df5732d1a737)|![image_6](https://github.com/user-attachments/assets/28378271-3782-484c-bd51-3d3311dd85c6)|
|<img src="https://github.com/user-attachments/assets/19d2abc4-36ae-4163-a8da-df5732d1a737" width="200"/> | <img src="https://github.com/user-attachments/assets/28378271-3782-484c-bd51-3d3311dd85c6" width="200"/> |
|-|-|
ControlNet 对于结构的控制力度是可以调节的,例如在下面这里例子中,我们把小姐姐从夏天移动到冬天时,适当调低 ControlNet 的控制力度,模型就会根据画面内容作出调整,为小姐姐换上温暖的衣服。
@@ -85,7 +84,7 @@ image_2 = pipe(
image_2.save("image_8.jpg")
```
|![image_7](https://github.com/user-attachments/assets/a7b8555b-bfd9-4e92-aa77-16bca81b07e3)|![image_8](https://github.com/user-attachments/assets/a1bab36b-6cce-4f29-8233-4cb824b524a8)|
|<img src="https://github.com/user-attachments/assets/a7b8555b-bfd9-4e92-aa77-16bca81b07e3" width="200"/> | <img src="https://github.com/user-attachments/assets/a1bab36b-6cce-4f29-8233-4cb824b524a8" width="200"/> |
|-|-|
## Upscaler/Tile/Blur: 高清图像生成
@@ -126,7 +125,7 @@ image_2 = pipe(
image_2.save("image_2.jpg")
```
|![image_1](https://github.com/user-attachments/assets/9038158a-118c-4ad7-ab01-22865f6a06fc)|![image_2](https://github.com/user-attachments/assets/88583a33-cd74-4cb9-8fd4-c6e14c0ada0c)|
|<img src="https://github.com/user-attachments/assets/9038158a-118c-4ad7-ab01-22865f6a06fc" width="200"/> | <img src="https://github.com/user-attachments/assets/88583a33-cd74-4cb9-8fd4-c6e14c0ada0c" width="200"/> |
|-|-|
```python
@@ -154,7 +153,7 @@ image_2 = pipe(
image_2.save("image_4.jpg")
```
|![image_3](https://github.com/user-attachments/assets/13061ecf-bb57-448a-82c6-7e4655c9cd85)|![image_4](https://github.com/user-attachments/assets/0b7ae80f-de58-4d1d-a49c-ad17e7631bdc)|
|<img src="https://github.com/user-attachments/assets/13061ecf-bb57-448a-82c6-7e4655c9cd85" width="200"/> | <img src="https://github.com/user-attachments/assets/0b7ae80f-de58-4d1d-a49c-ad17e7631bdc" width="200"/> |
|-|-|
## Inpaint: 局部重绘
@@ -196,7 +195,7 @@ image_2 = pipe(
image_2.save("image_10.jpg")
```
|![image_9](https://github.com/user-attachments/assets/babddad0-2d67-4624-b77a-c953250ebdab)|![mask_9](https://github.com/user-attachments/assets/d5bc2878-1817-457a-bdfa-200f955233d3)|![image_10](https://github.com/user-attachments/assets/e3197f2c-190b-4522-83ab-a2e0451b39f6)|
|<img src="https://github.com/user-attachments/assets/babddad0-2d67-4624-b77a-c953250ebdab" width="120"/> | <img src="https://github.com/user-attachments/assets/d5bc2878-1817-457a-bdfa-200f955233d3" width="120"/> |<img src="https://github.com/user-attachments/assets/e3197f2c-190b-4522-83ab-a2e0451b39f6" width="120"/> |
|-|-|-|
但是我们注意到,猫猫的头部动作发生了变化,如果我们想要保留原来的结构特征,可以使用 canny、depth、normal 模型DiffSynth-Studio 为不同结构的 ControlNet 提供了无缝的兼容支持。配合一个 normal ControlNet我们可以保证局部重绘时画面结构不变。
@@ -245,9 +244,10 @@ image_2 = pipe(
image_2.save("image_12.jpg")
```
|![image_11](https://github.com/user-attachments/assets/c028e6fc-5125-4cba-b35a-b6211c2e6600)|![mask_11](https://github.com/user-attachments/assets/1928ee9a-7594-4c6e-9c71-5bd0b043d8f4)|![image_12](https://github.com/user-attachments/assets/97b3b9e1-f821-405e-971b-9e1c31a209aa)|
|<img src="https://github.com/user-attachments/assets/c028e6fc-5125-4cba-b35a-b6211c2e6600" width="120"/> | <img src="https://github.com/user-attachments/assets/1928ee9a-7594-4c6e-9c71-5bd0b043d8f4" width="120"/> |<img src="https://github.com/user-attachments/assets/97b3b9e1-f821-405e-971b-9e1c31a209aa" width="120"/> |
|-|-|-|
## MultiControlNet+MultiDiffusion: 精细的高阶控制
DiffSynth-Studio 不仅支持多个不同结构的 ControlNet 同时生效,还支持使用不同提示词分区控制图中内容,还支持超高分辨率大图的分块处理,这让我们能够作出极为精细的高阶控制。接下来,我们展示一张精美图片的创作过程。