Lyra 2.0 Nvidia

Lyra 2.0 Nvidia is a framework for generating persistent, explorable 3D worlds at scale, designed for users who need to create simulation-ready 3D environments with interactive GUI. It addresses spatial forgetting and temporal drifting for long-horizon, 3D-consistent generation. The tool works by maintaining per-frame 3D geometry for information routing and using self-augmented training to teach the model to correct drift.
Lyra 2.0 Nvidia enables users to generate camera-controlled walkthrough videos that simulate scene walkthroughs, then lift them to 3D via feed-forward reconstruction techniques. This approach combines the visual fidelity and creative capacity of video models with 3D outputs ready for real-time rendering and simulation. The generated video can be further lifted into 3DGS and meshes, which can be directly exported to physics engines for downstream applications.
The tool is particularly valuable for users who need to create complex, interactive 3D environments, such as architects, game developers, and simulation engineers. These users can leverage Lyra 2.0 Nvidia to generate high-quality, 3D-consistent scenes that can be used for a variety of applications, including physically grounded robot navigation and interaction.