pasqualedem

/

label_anything_mae_480_coco

like 0

This PR significantly enhances the model card for "Label Anything: Multi-Class Few-Shot Semantic Segmentation with Visual Prompts" by adding essential metadata and comprehensive content.

Key improvements include:

• Adding license: mit and pipeline_tag: image-segmentation to the YAML metadata for better discoverability and categorization.
• Including the official paper abstract, project overview, and a visual demo image from the GitHub repository.
• Providing clear links to the Hugging Face paper page, project page, and GitHub repository.
• Incorporating "Quick Start" instructions with code snippets for one-line demo, manual installation, and model loading, making the model easier to use.
• Adding a table of pre-trained models and the official BibTeX citation.
• Removing redundant or internal information (like the run ID) from the original card.

Please review and merge if these updates align with the repository's needs!