Improve model card for Track-On2: Add pipeline tag, links, and comprehensive details
#1
by
nielsr
HF Staff
- opened
This PR significantly enhances the model card for the Track-On2 model by:
- Adding the
pipeline_tag: keypoint-detectionto the model metadata, improving discoverability on the Hugging Face Hub (e.g., at https://huggingface.co/models?pipeline_tag=keypoint-detection). - Prominently linking to the official paper: Track-On2: Enhancing Online Point Tracking with Memory.
- Including a direct link to the project page: https://kuis-ai.github.io/track_on2.
- Providing a link to the GitHub repository: https://github.com/gorkaydemir/track_on.
- Populating the model card content with an overview, installation instructions, sample usage code, details on pretrained models, datasets, evaluation, training, previous version information, and citation details, all adapted from the comprehensive GitHub README.
- Ensuring all image links and internal file references (like
demo.py) point to their correct raw URLs on the GitHub repository for proper rendering and access.
Please review and merge this PR to make this model more accessible and informative to the community.