This PR enhances the model card for **CoTAP: Semantic Concentration for Self-Supervised Dense Representations Learning**.
Key improvements include:
- Adding the `pipeline_tag: image-feature-extraction` to improve discoverability on the Hugging Face Hub (e.g., at https://huggingface.co/models?pipeline_tag=image-feature-extraction).
- Including `library_name: pytorch` as the model's code is based on PyTorch, enabling automated usage snippets.
- Integrating the full paper abstract to provide a comprehensive overview of the research.
- Linking to the official paper: [https://huggingface.co/papers/2509.09429](https://huggingface.co/papers/2509.09429).
- Linking to the official GitHub repository: [https://github.com/KID-7391/CoTAP](https://github.com/KID-7391/CoTAP).
- Incorporating detailed sections from the GitHub README, such as `Checkpoints`, `Environments`, `Installation`, `Training`, and `Evaluation`, to provide clear instructions for users.
- Embedding the pipeline overview image fro…
