# Pointcept

Pointcept: Perceive the world with sparse points, a codebase for point cloud perception research. Latest works: Utonia (ICML'26), Concerto (NeurIPS'25), Sonata (CVPR'25 Highlight), PTv3 (CVPR'24 Oral)

Repository: https://github.com/Pointcept/Pointcept
Canonical: https://ross.abutalabs.com/products/pointcept
Language: Python
License: MIT
License Family: permissive
Topics: 3d-vision, point-cloud, pytorch
Last push: 2026-08-03T04:44:30+00:00
Link (homepage): https://arxiv.org/abs/2306.03908
Link (site_page): https://info.arxiv.org/labs/index.html
Link (site_page): https://info.arxiv.org/about/ourmembers.html
Link (site_page): https://info.arxiv.org/about/donate.html
Link (site_page): https://info.arxiv.org/about

## Health v2 (maintenance only)
Score: 76/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 95, release rhythm 45, longevity 90
- inputs: {"age_days": 1261, "days_push": 30, "days_rel": 153, "gap_med": 186.0, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3196, forks 408 (observed 2026-08-28T04:07:48.344286+00:00)

## What it is
Pointcept is a PyTorch-based research codebase for point cloud perception, providing implementations of state-of-the-art 3D scene understanding models such as Point Transformer V3, Sonata, Concerto, and SAM3D. It offers training, evaluation, and pretrained weights for tasks like semantic and instance segmentation of 3D point clouds.

## Use cases
- train point cloud semantic segmentation models
- run 3D scene segmentation with Point Transformer V3
- apply Segment Anything to 3D point clouds
- benchmark point cloud perception models on ScanNet
- use pretrained 3D encoders for downstream tasks
- research self-supervised learning for point clouds

## When to choose
- you need state-of-the-art point cloud segmentation or backbone models
- you are doing 3D perception research in PyTorch
- you want pretrained 3D scene encoders like Sonata or PTv3

## When to avoid
- you need production-ready 3D inference services rather than research code
- your data is meshes or voxels rather than sparse point clouds
- you need a no-GPU or lightweight deployment

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, machine-learning, deep-learning
- platform: python
- tags: point-cloud, 3d-vision, pytorch, semantic-segmentation, scene-understanding, point-transformer, self-supervised-learning, research-codebase, research, gpu, linux

## Member repositories
- Pointcept/Pointcept (main) score 76
- Pointcept/PointTransformerV3 (plugin) score 48
- Pointcept/SegmentAnything3D (plugin) score 30

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:48.344286+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 2026-08-30T07:24:52.916640+00:00, confidence not recorded.
  - readme: https://github.com/Pointcept/Pointcept (fetched 2026-08-28T04:07:48.344286+00:00, sha f58b1ee5579f)
  - homepage: https://arxiv.org/abs/2306.03908 (fetched 2026-08-29T09:38:40.672828+00:00, sha 0dcbbd45f3b3)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T09:38:40.680545+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T09:38:40.675726+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T09:38:40.683241+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T09:38:40.677994+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.