# torch-points3d/torch-points3d

Pytorch framework for doing deep learning on point clouds.

Repository: https://github.com/torch-points3d/torch-points3d
Canonical: https://ross.abutalabs.com/products/torch-points3d
Homepage: https://torch-points3d.readthedocs.io/en/latest/
Language: Python
License: NOASSERTION
License Family: other
Topics: pytorch, deep-learning, point-cloud, pointnet, minkowskiengine, kpconv, segmentation, s3dis, scannet
Last push: 2026-08-24T21:08:13+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 8, longevity 100
- inputs: {"age_days": 2462, "days_push": 9, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2711, forks 404 (observed 2026-08-28T04:07:11.854492+00:00)

## What it is
A PyTorch-based framework for deep learning on 3D point clouds, supporting models like PointNet, KPConv, and MinkowskiEngine sparse convolutions. It provides a high-level API and Hydra-based configuration for reproducible training and benchmarking on datasets like S3DIS and ScanNet.

## Use cases
- train point cloud segmentation models
- benchmark deep learning models on 3D datasets
- run semantic segmentation on point clouds
- experiment with KPConv or PointNet architectures
- classify 3D point cloud data
- reproduce state-of-the-art point cloud benchmarks

## When to choose
- you need a high-level, configurable framework for point cloud deep learning in PyTorch
- you want reproducible benchmarking against standard 3D datasets like S3DIS or ScanNet
- you want to compare multiple point cloud architectures under one API

## When to avoid
- you work with 2D images or standard vision tasks rather than 3D point clouds
- you need a framework under active development with recent feature updates
- you cannot use CUDA or a sparse convolution backend

## Facets
- artifact type: framework
- maturity: maintenance
- function: deep-learning, machine-learning, computer-vision, benchmarking
- domain: deep-learning, computer-vision, machine-learning
- platform: python
- tags: point-cloud, pytorch, pointnet, kpconv, minkowskiengine, segmentation, 3d-vision, pytorch-geometric, gpu, docker, linux

## Member repositories
- torch-points3d/torch-points3d (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:11.854492+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 2026-08-30T02:15:35.265826+00:00, confidence not recorded.
  - readme: https://github.com/torch-points3d/torch-points3d (fetched 2026-08-28T04:07:11.854492+00:00, sha 1997b0196960)
- Data as of 2026-08-30T08:39:29.467469+00:00.