Improve model card: Add pipeline tag, library_name, paper, project page, code, and usage
#8
by nielsr HF Staff - opened
This PR significantly enhances the model card for the CVD-STORM model, addressing several missing pieces of information:
- Metadata:
- Adds
pipeline_tag: text-to-3dto categorize the model accurately on the Hub, reflecting its capability for generating videos with 4D (3D over time) reconstruction relevant to autonomous driving, including LiDAR generation. - Adds
library_name: diffusers, based on evidence from the GitHub README mentioning support for Stable Diffusion (SD 2.1, 3.0, 3.5) architectures, to enable easier programmatic access and the automated "how to use" widget.
- Adds
- Content:
- Updates the model card title to the paper title: CVD-STORM: Cross-View Video Diffusion with Spatial-Temporal Reconstruction Model for Autonomous Driving.
- Provides direct links to the paper, the project page (
https://sensetime-fvg.github.io/CVD-STORM/), and the GitHub repository (https://github.com/SenseTime-FVG/OpenDWM) for comprehensive information and usage examples. - Includes the abstract of the paper for quick understanding.
- Adds an "Installation" guide and a "Sample Usage" section with a code snippet, directly extracted from the GitHub README, to guide users on how to get started.
- Adds the citation information from the project's GitHub.
This update aims to make the model more discoverable, understandable, and usable for the community.