ComfyUI Node
Yogurt Sa2VA Video Segmentation
Segment an image batch or video with a reusable Sa2VA model loaded by Yogurt Sa2VA Model Loader.
Yogurt Sa2VA Video Segmentation
- sa2va_model
- images
- text_output
- masks
◄segmentation_promptPlease provide segmentation masks for the objects in this video.►
◄threshold0.50►
◄morphnone►
◄erode_kernel3►
◄dilate_kernel3►
◄iterations1►
CategoryYogurtSa2VA
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| sa2va_model | YOGURT_SA2VA_MODEL | — | |
| images | IMAGE | Batch of images or video frames to segment. | |
| segmentation_prompt | STRING | Please provide segmentation masks for the objects in this video. | Text prompt describing what objects to segment across all frames. |
| threshold | FLOAT | 0.500–1 | Threshold for converting probability masks to binary masks. |
| morph | COMBO | none | 5 options: none, opening, closing, erode, dilate |
| erode_kernel | INT | 31–50 | — |
| dilate_kernel | INT | 31–50 | — |
| iterations | INT | 11–10 | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| text_output | STRING | — |
| masks | MASK | — |