LoDAdaC: a unified local training-based decentralized framework with adaptive gradients and compressed communication

TL;DR

LoDAdaC is a decentralized learning framework that combines adaptive gradient methods with compressed communication to achieve faster convergence and reduced communication costs.

Contribution

It introduces a unified framework supporting multiple local training steps, adaptive optimizers, and compressed communication, with theoretical analysis and empirical validation.

Findings

LoDAdaC outperforms existing algorithms in convergence speed.

It significantly reduces communication costs through compression.

The framework is compatible with various adaptive optimizers.

Abstract

In the decentralized distributed learning, achieving fast convergence and low communication cost is essential for scalability and high efficiency. Adaptive gradient methods, such as Adam, have demonstrated strong practical performance in deep learning and centralized distributed settings. However, their convergence properties remain largely unexplored in decentralized settings involving multiple local training steps, such as federated learning. To address this limitation, we propose LoDAdaC, a unified multiple Local Training (MLT) Decentralized framework with Adam-type updates and Compressed communication (CC). LoDAdaC accommodates a broad class of optimizers for its local adaptive updates, including AMSGrad, Adam, and AdaGrad; it is compatible with standard (possibly biased) compressors such as low-bit quantization and sparsification. MLT and CC enable LoDAdaC to achieve multiplied reduction of communication cost, while the technique of adaptive updates enables fast convergence. We rigorously prove the combined advantage through complexity analysis. In addition, experiments on image classification and GPT-style language model training validate our theoretical findings and show that LoDAdaC significantly outperforms existing decentralized algorithms in terms of convergence speed and communication efficiency.