This PR significantly enhances the model card for **GS-Reasoner** by:
- Adding `pipeline_tag: image-text-to-text` to improve discoverability on the Hub for models performing 3D visual grounding and spatial reasoning.
- Specifying `library_name: transformers` to enable the automated "Use in Transformers" widget, as evidenced by `LlavaQwenForCausalLM` architecture, `LlavaProcessor`, `Qwen2Tokenizer`, and the usage of `LlavaAgent` within the project.
- Including a direct link to the paper on Hugging Face: [Reasoning in Space via Grounding in the World](https://huggingface.co/papers/2510.13800).
- Adding links to the project page ([https://yiming-cc.github.io/gs-reasoner/](https://yiming-cc.github.io/gs-reasoner/)) and the GitHub repository ([https://github.com/WU-CVGL/GS-Reasoner](https://github.com/WU-CVGL/GS-Reasoner)).
- Providing a detailed model description based on the paper's abstract.
- Including information about available model weights.
- Adding a "Sample Usage" s…
