This PR improves the model card by:
- Adding the `pipeline_tag: image-classification`, which will make the model discoverable at https://huggingface.co/models?pipeline_tag=image-classification.
- Adding a link to the paper: [DART: Differentiable Dynamic Adaptive Region Tokenizer for Vision Transformer and Mamba](https://huggingface.co/papers/2506.10390).
- Adding a link to the official GitHub repository: https://github.com/HCPLab-SYSU/DART.
Please review these changes.
