# open-mmlab/mmaction

An open-source toolbox for action understanding based on PyTorch

Repository: https://github.com/open-mmlab/mmaction
Canonical: https://ross.abutalabs.com/products/mmaction
Homepage: https://open-mmlab.github.io/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: action-recognition, action-detection, video-understanding, pytorch, temporal-action-detection, temporal-action-localization, spatial-temporal-action-detection
Last push: 2022-04-08T03:44:50+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2638, "days_push": 1608, "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 1876, forks 346 (observed 2026-08-28T04:05:47.773007+00:00)

## What it is
MMAction is an open-source PyTorch toolbox for video action understanding, covering action recognition, temporal action detection, and spatial-temporal action detection. It implements popular algorithms like TSN, I3D, SlowFast, R(2+1)D, and CSN with a modular design and pre-trained model zoo.

## Use cases
- recognize human actions in trimmed videos
- detect and localize actions in untrimmed videos
- train SlowFast or I3D models on Kinetics
- benchmark video action recognition algorithms
- run spatial-temporal action detection with pretrained models

## When to choose
- you need a PyTorch toolbox for video action recognition or detection
- you want reference implementations of TSN, I3D, SlowFast, R(2+1)D, or CSN
- you need pretrained video understanding models with a model zoo

## When to avoid
- you need actively maintained code - the repo has not had a release since 2022 and was superseded by MMAction2
- you need general video processing rather than action understanding
- you work outside PyTorch

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, video-processing, computer-vision
- domain: deep-learning, computer-vision, machine-learning
- platform: python
- tags: action-recognition, action-detection, video-understanding, pytorch, openmmlab, model-zoo, temporal-action-localization, video, linux, gpu, docker

## Member repositories
- open-mmlab/mmaction (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:47.773007+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 2026-08-30T03:14:44.783902+00:00, confidence not recorded.
  - readme: https://github.com/open-mmlab/mmaction (fetched 2026-08-28T04:05:47.773007+00:00, sha 70e38388ee29)
  - homepage: https://open-mmlab.github.io/ (fetched 2026-08-29T10:53:40.223186+00:00, sha 1cf9be975246)
- Data as of 2026-08-30T08:39:29.467469+00:00.