# kennymckormick/pyskl

A toolbox for skeleton-based action recognition.

Repository: https://github.com/kennymckormick/pyskl
Canonical: https://ross.abutalabs.com/products/pyskl
Language: Python
License: Apache-2.0
License Family: permissive
Topics: action-recognition, deep-learning, skeleton-based-action-recognition, stgcn, posec3d, gesture-recognition, pytorch
Last push: 2026-02-19T11:53:40+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 68, release rhythm 8, longevity 100
- inputs: {"age_days": 1614, "days_push": 195, "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 1261, forks 230 (observed 2026-08-28T04:04:10.066942+00:00)

## What it is
PYSKL is a PyTorch-based toolbox for skeleton-based human action recognition, built on MMAction2. It provides official implementations of PoseConv3D and ST-GCN++ along with training, testing, and skeleton extraction pipelines.

## Use cases
- recognize human actions from skeleton pose data
- train ST-GCN or PoseC3D models on NTU-RGB+D or Kinetics
- build a real-time gesture recognition demo
- extract skeletons from videos for action recognition
- benchmark skeleton-based action recognition algorithms
- train a two-stream model on RGB and pose data

## When to choose
- you need state-of-the-art skeleton-based action recognition models in PyTorch
- you want official PoseConv3D or ST-GCN++ implementations
- you work with pose/skeleton datasets like NTU-RGB+D

## When to avoid
- you need actively maintained software with timely bug fixes
- your task is RGB-video-only action recognition without skeletons
- you need a production-ready inference service out of the box

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision
- domain: deep-learning, computer-vision, machine-learning
- platform: python
- tags: action-recognition, skeleton-based, posec3d, stgcn, gesture-recognition, pytorch, mmaction2, linux, gpu

## Member repositories
- kennymckormick/pyskl (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:10.066942+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 2026-08-30T05:04:19.322983+00:00, confidence not recorded.
  - readme: https://github.com/kennymckormick/pyskl (fetched 2026-08-28T04:04:10.066942+00:00, sha 3abb15dd398c)
- Data as of 2026-08-30T08:39:29.467469+00:00.