Revisiting 3D ResNets for Video Recognition

Xianzhi Du, Yeqing Li, Yin Cui, Rui Qian, Jing Li +1 more
9/3/2021
cs.CVcs.LGeess.IV

Official code: tensorflow/models (77,659 stars, last commit 2026-04-29).

Abstract

A recent work from Bello shows that training and scaling strategies may be more significant than model architectures for visual recognition. This short note studies effective training and scaling strategies for video recognition models. We propose a simple scaling strategy for 3D ResNets, in combination with improved training strategies and minor architectural changes. The resulting models, termed 3D ResNet-RS, attain competitive performance of 81.0 on Kinetics-400 and 83.8 on Kinetics-600 without pre-training. When pre-trained on a large Web Video Text dataset, our best model achieves 83.5 and 84.3 on Kinetics-400 and Kinetics-600. The proposed scaling rule is further evaluated in a self-supervised setup using contrastive learning, demonstrating improved performance. Code is available at: https://github.com/tensorflow/models/tree/master/official.

View on arXivView PDF

Code Implementations(1)

tensorflow/modelsOfficial100%

Models and examples built with TensorFlow

77,65944,853C++, CSSFeb 5, 20165 months agoNOASSERTION

Similar Papers

Revisiting ResNets: Improved Training and Scaling Strategies

91% match1 repo5.3k2021

Delving Deeper into Data Scaling in Masked Image Modeling

87% match2023

Scalable Pre-training of Large Autoregressive Image Models

86% match1 repo1.4k2024

Swin Transformer V2: Scaling Up Capacity and Resolution

85% match1 repo16.0k2021

Masked Feature Prediction for Self-Supervised Visual Pre-Training

85% match1 repo7.4k2021

Scaling may be all you need for achieving human-level object recognition capacity with human-like visual experience

85% match1 repo142023

Cite this paper

@article{du2021revisiting,
  title  = {Revisiting 3D ResNets for Video Recognition},
  author = {Xianzhi Du and Yeqing Li and Yin Cui and Rui Qian and Jing Li and Irwan Bello},
  year   = {2021},
  eprint = {2109.01696},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2109.01696}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2109.01696.svg)](https://codeofpaper.com/paper/2109.01696)

Discussion