Creates a model card for the DiffusionSfM checkpoint (ckpt_00800000.pth), which previously had no README.
Changes:
- Add pipeline_tag: image-to-3d (multi-view images -> 3D scene geometry + camera poses)
- Add library_name: pytorch (raw .pth checkpoint, custom PyTorch stack; no transformers/diffusers artifacts)
- Add paper link to https://huggingface.co/papers/2505.05473 (CVPR 2025)
- Add license: mit (from the official GitHub repo LICENSE)
- Add brief model description, intended uses, and a minimal, verified usage snippet using huggingface_hub.hf_hub_download
- Link to the project page, GitHub repo, and Gradio demo Space
