DeepGlint-AI/mlcd-vit-base-patch32-224 · Improve model card: Add pipeline tag, library, abstract & usage

This PR significantly enhances the model card for rice-vit-large-patch14-560 by:

Adding pipeline_tag: image-feature-extraction for improved discoverability on the Hugging Face Hub.
Specifying library_name: transformers to indicate compatibility with the 🤗 Transformers library, enabling its "Use in Transformers" widget.
Updating the model's main title and incorporating the paper's abstract and highlights to provide a clearer and more detailed overview of the RICE method.
Including detailed usage examples and a dedicated section for visualizing semantic features, directly sourced from the original GitHub repository, making the model easier for users to interact with and understand.
Correcting the performance evaluation table header from "MLCD" to "RICE" to accurately reflect the reported results for this specific model.
Updating the paper link to the correct arXiv version for Region-based Cluster Discrimination (RICE) (https://arxiv.org/abs/2507.20025), as the previous link incorrectly pointed to a different paper.