55 lines
1.7 KiB
Markdown
55 lines
1.7 KiB
Markdown
|
|
# Documentation
|
|
- Class name: OneFormer-ADE20K-SemSegPreprocessor
|
|
- Category: ControlNet Preprocessors/Semantic Segmentation
|
|
- Output node: False
|
|
|
|
OneFormer ADE20K Segmentor节点专门用于语义分割任务，它使用在ADE20K数据集上预训练的OneFormer模型。该节点处理输入图像，对其进行语义分割，即根据ADE20K数据集的预定义类别，识别并分类图像中的每个像素。
|
|
|
|
# Input types
|
|
## Required
|
|
- image
|
|
- 需要进行语义分割的输入图像。该图像将由OneFormer模型处理，以识别和分类每个像素。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
## Optional
|
|
- resolution
|
|
- 在处理前将输入图像调整到的分辨率。这会影响分割输出的细节水平。
|
|
- Comfy dtype: INT
|
|
- Python dtype: int
|
|
|
|
# Output types
|
|
- image
|
|
- 语义分割过程的输出。这是一张图像，其中每个像素根据ADE20K数据集被分类到特定类别。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
|
|
|
|
## Usage tips
|
|
- Infra type: `GPU`
|
|
- Common nodes: unknown
|
|
|
|
|
|
## Source code
|
|
```python
|
|
class OneFormer_ADE20K_SemSegPreprocessor:
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return create_node_input_types()
|
|
|
|
RETURN_TYPES = ("IMAGE",)
|
|
FUNCTION = "semantic_segmentate"
|
|
|
|
CATEGORY = "ControlNet Preprocessors/Semantic Segmentation"
|
|
|
|
def semantic_segmentate(self, image, resolution=512):
|
|
from controlnet_aux.oneformer import OneformerSegmentor
|
|
|
|
model = OneformerSegmentor.from_pretrained(filename="250_16_swin_l_oneformer_ade20k_160k.pth")
|
|
model = model.to(model_management.get_torch_device())
|
|
out = common_annotator_call(model, image, resolution=resolution)
|
|
del model
|
|
return (out,)
|
|
|
|
```
|