# open-mmlab/mmtracking

OpenMMLab Video Perception Toolbox. It supports Video Object Detection (VID), Multiple Object Tracking (MOT), Single Object Tracking (SOT), Video Instance Segmentation (VIS) with a unified framework.

Repository: https://github.com/open-mmlab/mmtracking
Canonical: https://ross.abutalabs.com/products/mmtracking
Homepage: https://mmtracking.readthedocs.io/en/latest/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: single-object-tracking, video-object-detection, multi-object-tracking, video-instance-segmentation, tracking
Last push: 2023-09-19T07:31:38+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2195, "days_push": 1079, "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 3897, forks 621 (observed 2026-08-28T04:08:28.483699+00:00)

## What it is
MMTracking is OpenMMLab's PyTorch-based toolbox for video perception tasks, unifying video object detection, multiple object tracking, single object tracking, and video instance segmentation in one framework. It offers a modular design, pretrained model zoo, and integration with other OpenMMLab projects like MMDetection.

## Use cases
- track multiple objects across video frames
- run single object tracking on videos
- detect objects in video streams
- perform video instance segmentation
- train custom MOT models on my dataset
- benchmark state-of-the-art tracking models
- build a video analytics pipeline with pretrained detectors

## When to choose
- you need a unified PyTorch toolbox covering VID, MOT, SOT, and VIS
- you want pretrained state-of-the-art tracking models with a model zoo
- you already use OpenMMLab tools like MMDetection or MMPose
- you need modular components to compose custom tracking methods

## When to avoid
- you need real-time tracking on edge devices with minimal dependencies
- you want a actively developed project - it has been superseded by mmdetection 3.x and sees little activity
- you need non-PyTorch frameworks like TensorFlow or JAX
- you only need image-level detection without video/temporal modeling

## Facets
- artifact type: framework
- maturity: maintenance
- function: computer-vision, machine-learning, video-processing, deep-learning
- domain: computer-vision, machine-learning
- platform: python, cross-platform
- tags: object-tracking, multi-object-tracking, single-object-tracking, video-object-detection, video-instance-segmentation, pytorch, openmmlab, model-zoo, video, gpu, linux

## Member repositories
- open-mmlab/mmtracking (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:28.483699+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 2026-08-29T18:25:35.041310+00:00, confidence not recorded.
  - readme: https://github.com/open-mmlab/mmtracking (fetched 2026-08-28T04:08:28.483699+00:00, sha ca28ef2ca4a8)
- Data as of 2026-08-30T08:39:29.467469+00:00.