Zeroth-order fine-tuning of LLMs in random subspaces

Publication Type

Conference Proceeding Article

Publication Date

10-2025

Abstract

Fine-tuning Large Language Models (LLMs) has proven effective for a variety of downstream tasks. However, as LLMs grow in size, the memory demands for backpropagation become increasingly prohibitive. Zeroth-order (ZO) optimization methods offer a memory-efficient alternative by using forward passes to estimate gradients, but the variance of gradient estimates typically scales linearly with the model's parameter dimension–a significant issue for LLMs. In this paper, we propose the random Subspace Zeroth-order (SubZero) optimization to address the challenges posed by LLMs' high dimensionality. We introduce a low-rank perturbation tailored for LLMs that significantly reduces memory consumption while improving training performance. Additionally, we prove that our gradient estimation closely approximates the backpropagation gradient, exhibits lower variance than traditional ZO methods, and ensures convergence when combined with SGD. Experimental results show that SubZero enhances fine-tuning performance and achieves faster convergence compared to standard ZO approaches like MeZO across various language modeling tasks. Code is available at this https URL.

Discipline

Artificial Intelligence and Robotics

Research Areas

Intelligent Systems and Optimization

Publication

Proceedings of the 2025 International Conference on Computer Vision, ICCV, Honolulu, HawaiI, October 19-23

First Page

1

Last Page

27

Identifier

10.48550/arXiv.2410.08989

City or Country

Honolulu, HI, USA

Additional URL

https://doi.org/10.48550/arXiv.2410.08989

This document is currently not available here.

Share

COinS