Improve model card: Add metadata, links, and sample usage
#21
by
nielsr HF Staff - opened
This PR significantly enhances the model card for the CVD-STORM model (a component of the OpenDWM project).
It adds:
- The
library_name: diffusersmetadata, enabling integration with the Hugging Facediffuserslibrary based on the use of Stable Diffusion models. - The
pipeline_tag: text-to-3d, reflecting the model's capability to generate 3D/4D data like LiDAR from various conditions. - A link to the paper: CVD-STORM: Cross-View Video Diffusion with Spatial-Temporal Reconstruction Model for Autonomous Driving.
- A link to the official project page: https://sensetime-fvg.github.io/CVD-STORM/.
- A link to the GitHub repository: https://github.com/SenseTime-FVG/OpenDWM.
- A practical code snippet for layout-conditioned LiDAR generation using the Diffusion pipeline, directly taken from the GitHub README, to demonstrate usage.
Please review and merge if these improvements meet the community standards.