# 3D ResNets for Action Recognition

3D ResNets for Action Recognition (CVPR 2018)

Repository: https://github.com/kenshohara/3D-ResNets-PyTorch
Canonical: https://ross.abutalabs.com/products/3d-resnets-for-action-recognition
Language: Python
License: MIT
License Family: permissive
Topics: deep-learning, computer-vision, pytorch, python, action-recognition, video-recognition
Last push: 2021-01-20T20:31:36+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3275, "days_push": 2051, "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 4038, forks 926 (observed 2026-08-28T04:08:33.465767+00:00)

## What it is
A PyTorch implementation of 3D ResNet and R(2+1)D models for video action recognition, accompanying CVPR 2018 and related papers. It includes training, fine-tuning, and evaluation scripts plus pretrained models on Kinetics-700, Moments in Time, and STAIR-Actions.

## Use cases
- classify human actions in videos
- fine-tune a 3D CNN on my own video dataset
- get pretrained Kinetics video models
- train 3D ResNets on UCF-101 or HMDB-51
- extract spatiotemporal features from video clips
- reproduce CVPR 2018 action recognition results

## When to choose
- you need proven 3D CNN baselines for video action recognition
- you want pretrained Kinetics/Moments in Time models in PyTorch
- you are doing research on spatiotemporal video models

## When to avoid
- you need real-time or production video inference with minimal setup
- you want actively maintained code with recent PyTorch support
- you need lightweight 2D image classification instead of video

## Facets
- artifact type: library
- maturity: maintenance
- function: deep-learning, computer-vision, video-processing, machine-learning
- domain: computer-vision, deep-learning, machine-learning
- platform: python
- tags: pytorch, action-recognition, 3d-cnn, resnet, video-classification, pretrained-models, kinetics, video, linux, gpu

## Member repositories
- kenshohara/3D-ResNets-PyTorch (main) score 23
- kenshohara/video-classification-3d-cnn-pytorch (plugin) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:33.465767+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 2026-08-29T18:23:46.655128+00:00, confidence not recorded.
  - readme: https://github.com/kenshohara/3D-ResNets-PyTorch (fetched 2026-08-28T04:08:33.465767+00:00, sha d9ea4c43a4e1)
- Data as of 2026-08-30T08:39:29.467469+00:00.