This PR significantly improves the model card for UAGLNet by:
- Adding the `pipeline_tag: image-segmentation` to enhance discoverability on the Hugging Face Hub (e.g., at https://huggingface.co/models?pipeline_tag=image-segmentation).
- Specifying `library_name: pytorch`, based on evidence from the GitHub repository's installation instructions, which will enable an automated, predefined code snippet for PyTorch users.
- Including a direct link to the paper on Hugging Face: [https://huggingface.co/papers/2512.12941](https://huggingface.co/papers/2512.12941), alongside the existing arXiv link.
- Providing a concise description of the model based on the paper's abstract and GitHub README.
- Adding a "Sample Usage" section with code snippets directly from the GitHub README, demonstrating how to quickly reproduce results with pre-trained checkpoints.
- Removing the unnecessary "File information" section.
Please review and merge this PR if everything looks good.
