Skip to yearly menu bar Skip to main content


Poster

DoRA: Weight-Decomposed Low-Rank Adaptation

Shih-Yang Liu · Chien-Yi Wang · Hongxu Yin · Pavlo Molchanov · Yu-Chiang Wang · Kwang-Ting Cheng · Min-Hung Chen

Hall C 4-9 #607
[ ] [ Project Page ] [ Paper PDF ]
Thu 25 Jul 4:30 a.m. PDT — 6 a.m. PDT
 
Oral presentation: Oral 6B Low Rank Learning
Thu 25 Jul 7:30 a.m. PDT — 8:30 a.m. PDT

Abstract:

Among the widely used parameter-efficient fine-tuning (PEFT) methods, LoRA and its variants have gained considerable popularity because of avoiding additional inference costs. However, there still often exists an accuracy gap between these methods and full fine-tuning (FT). In this work, we first introduce a novel weight decomposition analysis to investigate the inherent differences between FT and LoRA. Aiming to resemble the learning capacity of FT from the findings, we propose Weight-Decomposed Low-Rank Adaptation (DoRA). DoRA decomposes the pre-trained weight into two components, magnitude and direction, for fine-tuning, specifically employing LoRA for directional updates to efficiently minimize the number of trainable parameters. By employing DoRA, we enhance both the learning capacity and training stability of LoRA while avoiding any additional inference overhead. DoRA consistently outperforms LoRA on fine-tuning LLaMA, LLaVA, and VL-BART on various downstream tasks, such as commonsense reasoning, visual instruction tuning, and image/video-text understanding. The code is available at https://github.com/NVlabs/DoRA.

Live content is unavailable. Log in and register to view live content