This PR significantly enhances the model card for the `OpenGVLab/Hulk` model.
It includes:
- Adding `pipeline_tag: any-to-any` to classify the model's generalist, multimodal capabilities.
- Adding `library_name: transformers` to indicate compatibility with the Transformers library and enable the "Use in Transformers" widget.
- Adding comprehensive tags like `human-centric`, `multimodal`, `2d-vision`, `3d-vision`, `vision-language`, and various task-specific tags for better discoverability.
- Incorporating the full paper abstract, project page, and GitHub repository links for detailed information.
- Including key visuals (teaser and framework images) from the GitHub repository.
- Providing clear guidance on where to find usage examples, training, and evaluation details (the official GitHub repository).
- Adding the academic citation.
These changes will greatly improve the model's visibility and utility on the Hugging Face Hub.
Upload the config of Hulk(ViT-B), which is similar to https://github.com/OpenGVLab/Hulk/blob/main/experiments/release/Hulk_vit-B.yaml