Add model card

This PR adds a model card for GSFixer, linking it to the paper [GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors](https://huggingface.co/papers/2508.09667) and the official GitHub repository. It also adds relevant metadata for the `diffusers` library and the `image-to-video` pipeline tag based on the model's configuration.

README.md

@@ -0,0 +1,28 @@
+---
+library_name: diffusers
+pipeline_tag: image-to-video
+---
+
+# GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors
+
+GSFixer is a framework designed to improve the quality of 3D Gaussian Splatting (3DGS) representations reconstructed from sparse inputs. It leverages a reference-guided video restoration model built upon a DiT-based video diffusion model (CogVideoX).
+
+- **Paper:** [GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors](https://huggingface.co/papers/2508.09667)
+- **GitHub:** [GVCLab/GSFixer](https://github.com/GVCLab/GSFixer)
+
+## Introduction
+
+Reconstructing 3D scenes using 3D Gaussian Splatting (3DGS) from sparse views often results in noticeable artifacts due to insufficient information. GSFixer addresses this challenge by using a reference-guided video restoration approach. By integrating both 2D semantic features (from DINOv2) and 3D geometric features (from VGGT) of reference views, the model enhances semantic coherence and 3D consistency when fixing novel views rendered from low-quality 3DGS.
+
+## Citation
+
+If you find this work useful, please consider citing:
+
+```bibtex
+@article{yin2025gsfixer,
+  title={GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors},
+  author={Yin, Xingyilang and Zhang, Qi and Chang, Jiahao and Feng, Ying and Fan, Qingnan and Yang, Xi and Pun, Chi-Man and Zhang, Huaqi and Cun, Xiaodong},
+  journal={arXiv preprint arXiv:2508.09667},
+  year={2025}
+}
+```