arXiv · 2410.19694
Less is More: Extreme Gradient Boost Rank-1 Adaption for Efficient Finetuning of LLMs
Abstract
Fine-tuning Large Language Models (LLMs) has become a crucial technique for adapting pre-trained models to downstream tasks. However, the enormous size of LLMs poses significant challenges in terms of computational complexity and resource requirements. Low-Rank Adaptation (LoRA) has emerged as a promising solution. However, there exists a gap between the practical performance of low-rank adaptations and its theoretical optimum. In this work, we propose eXtreme Gradient Boosting LoRA (XGBLoRA), a novel framework that bridges this gap by leveraging the power of ensemble learning. Inspired by gradient boosting, XGBLoRA iteratively learns and merges a sequence of LoRA adaptations to refine model predictions. It achieves better performance than the standard LoRA, while enjoying the computational efficiency of rank-1 adaptations. We provide theoretical analysis to show the convergence and optimality of our approach, and conduct extensive experiments on a range of natural language processing tasks. The results demonstrate that XGBLoRA consistently outperforms standard LoRA and achieves performance comparable to full fine-tuning with significantly fewer trainable parameters. This work advances parameter-efficient fine-tuning for LLMs, and offers a promising solution for adapting LLMs to downstream tasks while optimizing performance and efficiency.
Explore related subjects
Keep this discovery
Yifei Zhang, Hao Zhu, Aiwei Liu, Han Yu, Piotr Koniusz, Irwin King. 2024-10-25. Less is More: Extreme Gradient Boost Rank-1 Adaption for Efficient Finetuning of LLMs. https://arxiv.org/abs/2410.19694
Cite the original work for its findings. Save a collection to share your selection of sources.