# traveller59/second.pytorch

SECOND for KITTI/NuScenes object detection

Repository: https://github.com/traveller59/second.pytorch
Canonical: https://ross.abutalabs.com/products/secondpytorch
Language: Python
License: MIT
License Family: permissive
Topics: kitti, object-detection, voxelnet, nuscenes
Last push: 2022-10-14T08:05:18+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2891, "days_push": 1419, "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 1777, forks 709 (observed 2026-08-28T04:05:34.660145+00:00)

## What it is
A PyTorch implementation of the SECOND 3D object detector for LiDAR point clouds, supporting KITTI and NuScenes datasets with sparse convolution-based networks. The project is deprecated in favor of OpenPCDet and mmdetection3d.

## Use cases
- train a 3D object detection model on KITTI point clouds
- run SECOND detector on NuScenes dataset
- detect cars and pedestrians from LiDAR scans
- train PointPillars for autonomous driving perception
- benchmark 3D detection AP on KITTI validation set
- prototype voxel-based sparse convolution detectors

## When to choose
- you need the original SECOND reference implementation for research comparison
- you must reproduce legacy KITTI/NuScenes results from 2019-era papers
- you are working with spconv 1.x codebases

## When to avoid
- you are starting a new project - use OpenPCDet or mmdetection3d instead
- you need spconv 2.x support or active maintenance
- you need production-ready 3D detection with bug fixes

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, deep-learning
- domain: computer-vision, autonomous-vehicles, deep-learning
- platform: windows, python
- tags: object-detection, lidar, point-clouds, kitti, nuscenes, voxelnet, pointpillars, spconv, pytorch, deprecated, linux, gpu

## Member repositories
- traveller59/second.pytorch (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:34.660145+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 2026-08-30T03:25:14.663958+00:00, confidence not recorded.
  - readme: https://github.com/traveller59/second.pytorch (fetched 2026-08-28T04:05:34.660145+00:00, sha ecaa12ff455c)
- Data as of 2026-08-30T08:39:29.467469+00:00.