# nv-tlabs/lyra

Project Lyra: Open Generative 3D World Models

Repository: https://github.com/nv-tlabs/lyra
Canonical: https://ross.abutalabs.com/products/nv-tlabs-lyra
Homepage: https://research.nvidia.com/labs/sil/projects/lyra2/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: 3d, diffusion, gaussians, video, world-model
Last push: 2026-07-20T17:07:42+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 93, release rhythm 35, longevity 25
- inputs: {"age_days": 358, "days_push": 44, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2262, forks 229 (observed 2026-08-28T04:06:31.936835+00:00)

## What it is
Project Lyra is NVIDIA's open series of generative 3D world models, including Lyra 1.0 for feed-forward 3D/4D scene generation from a single image or video and Lyra 2.0 for explorable, long-horizon 3D-consistent world generation. It provides official model weights, training code, and a GUI.

## Use cases
- generate 3D scenes from a single image
- create explorable 3D worlds from video
- reconstruct 3D scenes with gaussian splatting
- train a generative 3D world model
- generate 4D scenes from video
- run a 3D world model locally with a GUI

## When to choose
- you need state-of-the-art generative 3D scene or world generation
- you want to build interactive, explorable 3D environments from images or video
- you are researching video-diffusion-based 3D reconstruction and have GPU resources

## When to avoid
- you need lightweight 3D modeling without heavy GPU compute
- you need a production game engine rather than a research model
- you require permissive licensing of the model weights, not just the Apache-2.0 code

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, video-processing, graphics, simulation
- domain: artificial-intelligence, computer-vision, graphics, deep-learning, machine-learning
- platform: python
- tags: 3d-generation, world-models, gaussian-splatting, diffusion-models, scene-reconstruction, generative-ai, nvidia, linux, gpu

## Member repositories
- nv-tlabs/lyra (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:31.936835+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:42:32.041026+00:00, confidence not recorded.
  - readme: https://github.com/nv-tlabs/lyra (fetched 2026-08-28T04:06:31.936835+00:00, sha a23452885a1a)
  - homepage: https://research.nvidia.com/labs/sil/projects/lyra2/ (fetched 2026-08-29T10:23:03.073936+00:00, sha d1c5bdd55b4b)
- Data as of 2026-08-30T08:39:29.467469+00:00.