Add model card for Lyra 2.0

This PR populates the model card for Lyra 2.0 with relevant information from the paper and official repository. It includes:
- Metadata for the `image-to-3d` pipeline and the Apache 2.0 license.
- Links to the paper ([2604.13036](https://huggingface.co/papers/2604.13036)), project page, and official GitHub repository.
- A brief summary of the framework's approach to long-horizon, 3D-consistent scene generation.
- Citation information.

README.md

@@ -1,5 +1,36 @@
 ---
-{}
+license: apache-2.0
+pipeline_tag: image-to-3d
 ---
 
-Model Release Today. Stay tuned!
+# Lyra 2.0: Explorable Generative 3D Worlds
+
+[Paper](https://huggingface.co/papers/2604.13036) | [Project Page](https://research.nvidia.com/labs/sil/lyra2/) | [Code](https://github.com/nv-tlabs/lyra)
+
+Lyra 2.0 is a framework for generating persistent, explorable 3D worlds at scale. It addresses the challenges of spatial forgetting and temporal drifting in long-horizon 3D scene creation. By maintaining per-frame 3D geometry for information routing and using self-augmented histories during training to correct synthesis errors, Lyra 2.0 enables substantially longer and 3D-consistent video trajectories that can be reconstructed into high-quality 3D scenes.
+
+## Overview
+
+Recent advances in video generation enable a new paradigm for 3D scene creation: generating camera-controlled videos that simulate scene walkthroughs, then lifting them to 3D via feed-forward reconstruction techniques. Lyra 2.0 scales this approach to large, complex environments through two key mechanisms:
+- **Geometry-aware Information Routing**: Maintaining per-frame 3D geometry to retrieve relevant past frames and establish correspondences.
+- **Drift Correction**: Training with self-augmented histories to teach the model to correct accumulated synthesis errors rather than propagate them.
+
+## Citation
+
+```bibtex
+@article{shen2026lyra2,
+  title={Lyra 2.0: Explorable Generative 3D Worlds},
+  author={Shen, Tianchang and Bahmani, Sherwin and He, Kai and Srinivasan, Sangeetha Grama and Cao, Tianshi and Ren, Jiawei and Li, Ruilong and Wang, Zian and Sharp, Nicholas and Gojcic, Zan and Fidler, Sanja and Huang, Jiahui and Ling, Huan and Gao, Jun and Ren, Xuanchi},
+  journal={arXiv preprint arXiv:2604.13036},
+  year={2026}
+}
+
+@inproceedings{bahmani2026lyra,
+  title={Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation},
+  author={Bahmani, Sherwin and Shen, Tianchang and Ren, Jiawei and Huang, Jiahui and Jiang, Yifeng and 
+          Turki, Haithem and Tagliasacchi, Andrea and Lindell, David B. and Gojcic, Zan and Fidler, Sanja and 
+          Ling, Huan and Gao, Jun and Ren, Xuanchi},
+  booktitle={International Conference on Learning Representations (ICLR)},
+  year={2026}
+}
+```