# Tencent/ObjectDetection-OneStageDet

单阶段通用目标检测器

Repository: https://github.com/Tencent/ObjectDetection-OneStageDet
Canonical: https://ross.abutalabs.com/products/objectdetection-onestagedet
Language: Python
License: NOASSERTION
License Family: other
Last push: 2025-07-02T06:11:06+00:00

## Health v2 (maintenance only)
Score: 45/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 29, release rhythm 35, longevity 100
- inputs: {"age_days": 2814, "days_push": 427, "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 1980, forks 498 (observed 2026-08-28T04:06:01.686631+00:00)

## What it is
OneStageDet (OSD) is a PyTorch-based single-stage object detection framework from Tencent implementing YOLOv2 and YOLOv3 with plans to unify YOLO and SSD in one framework. It includes many efficient backbones (tiny YOLO, MobileNet, ShuffleNet, SqueezeNext, Xception) and benchmarking tools for detection speed.

## Use cases
- train a custom YOLOv3 object detector in PyTorch
- benchmark inference speed of detection networks
- experiment with lightweight backbones like MobileNet or ShuffleNet for detection
- reproduce YOLOv2/YOLOv3 results on Pascal VOC
- train tiny YOLO models from scratch for edge deployment

## When to choose
- you want a readable PyTorch implementation of YOLOv2/YOLOv3 to study or modify
- you need to compare many efficient backbones within one detection framework
- you want training speed comparable to darknet but in Python

## When to avoid
- you need a production-grade, actively maintained detector with modern architectures (e.g., YOLOv8, DETR)
- you require multi-stage detectors or the latest SOTA accuracy
- you need Windows support or recent PyTorch compatibility out of the box

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, computer-vision, image-processing, benchmarking
- domain: computer-vision, deep-learning, machine-learning, image-processing
- platform: python
- tags: object-detection, yolo, pytorch, single-stage-detector, yolov3, yolov2, ssd, deep-learning-framework, linux, gpu

## Member repositories
- Tencent/ObjectDetection-OneStageDet (main) score 45

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:01.686631+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 2026-08-30T03:03:59.709043+00:00, confidence not recorded.
  - readme: https://github.com/Tencent/ObjectDetection-OneStageDet (fetched 2026-08-28T04:06:01.686631+00:00, sha 11068b56805c)
- Data as of 2026-08-30T08:39:29.467469+00:00.