# ultralytics/yolov3

PyTorch implementation of YOLOv3, YOLOv3-SPP, and YOLOv3-tiny for real-time object detection with training, validation, inference, and multi-format export.

Repository: https://github.com/ultralytics/yolov3
Canonical: https://ross.abutalabs.com/products/yolov3
Homepage: https://docs.ultralytics.com/models/yolov3/
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: yolov3, object-detection, yolo, deep-learning, ultralytics, computer-vision, coreml, edge-ai, inference, model-export, model-training, onnx, pytorch, tensorrt
Last push: 2026-08-20T20:13:23+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 8, longevity 100
- inputs: {"age_days": 2929, "days_push": 13, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 10596, forks 3429 (observed 2026-08-28T04:10:42.504469+00:00)

## What it is
Ultralytics' PyTorch implementation of YOLOv3, YOLOv3-SPP, and YOLOv3-tiny for real-time object detection. It provides training, validation, inference, and multi-format model export (ONNX, TensorRT, CoreML) tooling.

## Use cases
- detect objects in images and videos in real time
- train a YOLOv3 model on a custom dataset
- run object detection inference on GPU or edge devices
- export a detection model to ONNX, TensorRT, or CoreML
- validate object detection accuracy on COCO-style datasets
- deploy a lightweight detector with YOLOv3-tiny

## When to choose
- you specifically need the classic YOLOv3 architecture or its tiny/SPP variants
- you want a mature, well-documented PyTorch detection codebase with export tooling
- you need real-time detection on resource-constrained edge hardware

## When to avoid
- you want the latest Ultralytics models and features - use the ultralytics package with YOLO11/YOLO26 instead
- you need segmentation, pose estimation, or tracking rather than plain detection
- you require a permissive license - AGPL-3.0 restricts commercial use without an enterprise license

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, machine-learning, artificial-intelligence
- platform: python, cross-platform
- tags: yolov3, object-detection, pytorch, real-time-detection, model-training, onnx, tensorrt, coreml, edge-ai, model-export, gpu, docker

## Member repositories
- ultralytics/yolov3 (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:42.504469+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 2026-08-29T17:18:29.733409+00:00, confidence not recorded.
  - readme: https://github.com/ultralytics/yolov3 (fetched 2026-08-28T04:10:42.504469+00:00, sha acd37f33c405)
  - homepage: https://docs.ultralytics.com/models/yolov3/ (fetched 2026-08-29T08:17:40.770872+00:00, sha bc6e3fceaedc)
  - site_page: https://docs.ultralytics.com/quickstart (fetched 2026-08-29T08:17:40.780503+00:00, sha 1c588270a735)
  - site_page: https://docs.ultralytics.com/integrations (fetched 2026-08-29T08:17:40.782752+00:00, sha a82fee1740bf)
  - site_page: https://docs.ultralytics.com/ (fetched 2026-08-29T08:17:40.784996+00:00, sha fba5847c3417)
  - site_page: https://docs.ultralytics.com/modes/train (fetched 2026-08-29T08:17:40.787105+00:00, sha 5f6f930f49db)
  - site_page: https://docs.ultralytics.com/modes/predict (fetched 2026-08-29T08:17:40.789968+00:00, sha cf4c377dca7d)
  - site_page: https://docs.ultralytics.com/tasks/detect (fetched 2026-08-29T08:17:40.792666+00:00, sha ac9cb5538b5f)
- Data as of 2026-08-30T08:39:29.467469+00:00.