FALQON: Accelerating LoRA Fine-tuning with Low-Bit Floating-Point Arithmetic

Kanghyun Choi, Hyeyoon Lee, SunJong Park, Dain Kwon, Jinho Lee

Keywords: FP8, Quantization, LLM, LoRA, Quantized training

TL;DR: FALQON accelerates LoRA fine-tuning by up to 3$\times$ through merging adapters into an FP8-quantized backbone, removing redundant quantization overhead from small matrices.

Abstract: Low-bit floating-point (FP) formats, such as FP8, provide significant acceleration and memory savings in model training thanks to native hardware support on modern GPUs and NPUs. However, we analyze that FP8 quantization offers speedup primarily for large-dimensional matrix multiplications, while inherent quantization overheads diminish speedup when applied to low-rank adaptation (LoRA), which uses small-dimensional matrices for efficient fine-tuning of large language models (LLMs). To address this limitation, we propose FALQON, a novel framework that eliminates the quantization overhead from separate LoRA computational paths by directly merging LoRA adapters into an FP8-quantized backbone during fine-tuning. Furthermore, we reformulate the forward and backward computations for merged adapters to significantly reduce quantization overhead, and introduce a row-wise proxy update mechanism that efficiently integrates substantial updates into the quantized backbone. Experimental evaluations demonstrate that FALQON achieves approximately a 3$\times$ training speedup over existing quantized LoRA methods with a similar level of accuracy, providing a practical solution for efficient large-scale model fine-tuning. Moreover, FALQON’s end-to-end FP8 workflow removes the need for post-training quantization, facilitating efficient deployment. Code is available at https://github.com/iamkanghyunchoi/falqon.

Primary Area: Deep learning (e.g., architectures, generative models, optimization for deep networks, foundation models, LLMs)

Submission Number: 5149