FedMomentum: Preserving LoRA Training Momentum in Federated Fine-Tuning

TL;DR

FedMomentum introduces a SVD-based aggregation method for federated LoRA fine-tuning of LLMs, effectively preserving training momentum, improving convergence speed, and enhancing task-specific performance.

Contribution

The paper proposes FedMomentum, a novel structured aggregation framework that maintains LoRA training momentum using SVD, addressing structural and convergence issues in federated fine-tuning.

Findings

FedMomentum outperforms existing methods in convergence speed.

It achieves higher final accuracy across multiple tasks.

The approach effectively preserves LoRA's low-rank structure.

Abstract

Federated fine-tuning of large language models (LLMs) with low-rank adaptation (LoRA) offers a communication-efficient and privacy-preserving solution for task-specific adaptation. Naive aggregation of LoRA modules introduces noise due to mathematical incorrectness when averaging the downsampling and upsampling matrices independently. However, existing noise-free aggregation strategies inevitably compromise the structural expressiveness of LoRA, limiting its ability to retain client-specific adaptations by either improperly reconstructing the low-rank structure or excluding partially trainable components. We identify this problem as loss of training momentum, where LoRA updates fail to accumulate effectively across rounds, resulting in slower convergence and suboptimal performance. To address this, we propose FedMomentum, a novel framework that enables structured and momentum-preserving LoRA aggregation via singular value decomposition (SVD). Specifically, after aggregating low-rank updates in a mathematically correct manner, FedMomentum applies SVD to extract the dominant components that capture the main update directions. These components are used to reconstruct the LoRA modules with the same rank, while residual components can be retained and later merged into the backbone to preserve semantic information and ensure robustness. Extensive experiments across multiple tasks demonstrate that FedMomentum consistently outperforms prior state-of-the-art methods in convergence speed and final accuracy.