🦖 Talk2DINO & Talk2DINOv3 Demo

Perform open-vocabulary semantic segmentation using Talk2DINO or Talk2DINOv3.
Supports both ViT-B and ViT-L model sizes.

Steps:

1. Upload or select an example image.
2. Enter class names (e.g., pikachu, forest, road).
3. Choose model variant (Talk2DINO or v3).
4. Adjust options and click Generate Segmentation.