# NVlabs/VoxFormer

Official PyTorch implementation of VoxFormer [CVPR 2023 Highlight]

Repository: https://github.com/NVlabs/VoxFormer
Canonical: https://ross.abutalabs.com/products/voxformer
Language: Python
License: NOASSERTION
License Family: other
Topics: 3d-scene-understanding, artificial-intelligence, autonomous-driving, autonomous-vehicles, computer-vision, semantic-scene-completion, vision-transformer, 3d-perception, occupancy-grid-map, machine-learning, voxel-proceessing, 2d-to-3d, deep-learning, semantickitti
Last push: 2023-12-07T22:46:17+00:00

## Health v2 (maintenance only)
Score: 31/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 92
- inputs: {"age_days": 1289, "days_push": 1000, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1208, forks 102 (observed 2026-08-28T04:03:59.611420+00:00)

## What it is
Official PyTorch implementation of VoxFormer, a CVPR 2023 highlight paper presenting a sparse voxel transformer for camera-based 3D semantic scene completion. It predicts complete 3D volumetric semantics from 2D images via a two-stage sparse-to-dense design, achieving state-of-the-art results on SemanticKITTI.

## Use cases
- predict 3d semantic occupancy maps from camera images
- run semantic scene completion on semantickitti
- research camera-based 3d perception for autonomous driving
- compare 3d scene completion baselines
- reproduce cvpr 2023 voxel transformer results
- generate occupancy grid maps from monocular or stereo cameras

## When to choose
- you need a strong baseline for camera-only 3D semantic scene completion
- you are researching occupancy prediction or voxel transformers for autonomous driving
- you want to benchmark on SemanticKITTI or SSCBench

## When to avoid
- you need LiDAR-based scene completion rather than camera-only
- you need a production-ready perception stack rather than research code
- you lack a GPU or PyTorch training environment

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, autonomous-vehicles, artificial-intelligence, deep-learning
- platform: python
- tags: semantic-scene-completion, 3d-occupancy-prediction, voxel-transformer, pytorch, semantickitti, camera-based-perception, cvpr-2023, research-code, linux, gpu

## Member repositories
- NVlabs/VoxFormer (main) score 31

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:59.611420+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 2026-08-30T06:18:38.469092+00:00, confidence not recorded.
  - readme: https://github.com/NVlabs/VoxFormer (fetched 2026-08-28T04:03:59.611420+00:00, sha b5913860a797)
- Data as of 2026-08-30T08:39:29.467469+00:00.