Among the widely used parameter-efficient finetuning (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 LowRank 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.

翻译：在广泛使用的参数高效微调（PEFT）方法中，LoRA及其变体因避免了额外推理成本而备受青睐。然而，这些方法与全参数微调（FT）之间仍常存在精度差距。本文首先提出一种新颖的权重分解分析方法，以探究FT与LoRA之间的本质差异。旨在从研究发现中模拟FT的学习能力，我们提出了权重分解的低秩适配（DoRA）。DoRA将预训练权重分解为幅度和方向两个分量进行微调，特别采用LoRA进行方向性更新，以高效最小化可训练参数数量。通过采用DoRA，我们在提升LoRA学习能力和训练稳定性的同时，避免了任何额外的推理开销。在LLaMA、LLaVA和VL-BART的微调任务中，DoRA在常识推理、视觉指令调优以及图像/视频文本理解等多项下游任务上均持续优于LoRA。