# Quasi-Global Momentum: Accelerating Decentralized Deep Learning on Heterogeneous Data

Source: [Quasi-Global Momentum: Accelerating Decentralized Deep Learning on Heterogeneous Data](<https://www.sota2.com/research/paper/68fa6a3d1e209131cea26e93>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T23:19:37.191Z

Authors: Tao Lin, Sai Praneeth Karimireddy, Sebastian U. Stich, Martin Jaggi

Published: 2021-02-09T11:27:14.000Z

## Abstract

Decentralized training of deep learning models is a key element for enabling data privacy and on-device learning over networks. In realistic learning scenarios, the presence of heterogeneity across different clients' local datasets poses an optimization challenge and may severely deteriorate the generalization performance. In this paper, we investigate and identify the limitation of several decentralized optimization algorithms for different degrees of data heterogeneity. We propose a novel momentum-based method to mitigate this decentralized training difficulty. We show in extensive empirical experiments on various CV/NLP datasets (CIFAR-10, ImageNet, and AG News) and several network topologies (Ring and Social Network) that our method is much more robust to the heterogeneity of clients' data than other existing methods, by a significant improvement in test performance ($1\\% \\!-\\! 20\\%$). Our code is publicly available.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2102.04761>)

[PDF](<https://arxiv.org/pdf/2102.04761.pdf>)