@inproceedings{wang-etal-2024-prolora,
title = "{PR}o{L}o{RA}: Partial Rotation Empowers More Parameter-Efficient {L}o{RA}",
author = "Wang, Sheng and
Xue, Boyang and
Ye, Jiacheng and
Jiang, Jiyue and
Chen, Liheng and
Kong, Lingpeng and
Wu, Chuan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = aug,
year = "2024",
address = "Bangkok, Thailand",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.acl-long.156",
pages = "2829--2841",
abstract = "With the rapid scaling of large language models (LLMs), serving numerouslow-rank adaptations (LoRAs) concurrently has become increasingly impractical,leading to unaffordable costs and necessitating more parameter-efficientfinetuning methods. In this work, we introduce $\text{\textbf{P}artially \textbf{Ro}tation-enhanced \textbf{Lo}w-\textbf{R}ank \textbf{A}daptation (PRoLoRA)}$, an intra-layer sharing mechanism comprising fouressential components: broadcast reduction, rotation enhancement,partially-sharing refinement, and rectified initialization strategy. As asuperset of LoRA, PRoLoRA retains its advantages, and effectively circumventthe drawbacks of peer parameter-sharing methods with superior model capacity,practical feasibility, and broad applicability. Empirical experimentsdemonstrate the remarkably higher parameter efficiency of PRoLoRA in bothspecific parameter budget and performance target scenarios, and its scalabilityto larger LLMs. Notably, with one time less trainable parameters, PRoLoRA stilloutperforms LoRA on multiple instruction tuning datasets. Subsequently, anablation study is conducted to validate the necessity of individual componentsand highlight the superiority of PRoLoRA over three potential variants.Hopefully, the conspicuously higher parameter efficiency can establish PRoLoRAas a resource-friendly alternative to LoRA.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2024-prolora">
<titleInfo>
<title>PRoLoRA: Partial Rotation Empowers More Parameter-Efficient LoRA</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sheng</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Boyang</namePart>
<namePart type="family">Xue</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiacheng</namePart>
<namePart type="family">Ye</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiyue</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liheng</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lingpeng</namePart>
<namePart type="family">Kong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chuan</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lun-Wei</namePart>
<namePart type="family">Ku</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Martins</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="family">Srikumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Bangkok, Thailand</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>With the rapid scaling of large language models (LLMs), serving numerouslow-rank adaptations (LoRAs) concurrently has become increasingly impractical,leading to unaffordable costs and necessitating more parameter-efficientfinetuning methods. In this work, we introduce \textPartially Rotation-enhanced Low-Rank Adaptation (PRoLoRA), an intra-layer sharing mechanism comprising fouressential components: broadcast reduction, rotation enhancement,partially-sharing refinement, and rectified initialization strategy. As asuperset of LoRA, PRoLoRA retains its advantages, and effectively circumventthe drawbacks of peer parameter-sharing methods with superior model capacity,practical feasibility, and broad applicability. Empirical experimentsdemonstrate the remarkably higher parameter efficiency of PRoLoRA in bothspecific parameter budget and performance target scenarios, and its scalabilityto larger LLMs. Notably, with one time less trainable parameters, PRoLoRA stilloutperforms LoRA on multiple instruction tuning datasets. Subsequently, anablation study is conducted to validate the necessity of individual componentsand highlight the superiority of PRoLoRA over three potential variants.Hopefully, the conspicuously higher parameter efficiency can establish PRoLoRAas a resource-friendly alternative to LoRA.</abstract>
<identifier type="citekey">wang-etal-2024-prolora</identifier>
<location>
<url>https://aclanthology.org/2024.acl-long.156</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>2829</start>
<end>2841</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PRoLoRA: Partial Rotation Empowers More Parameter-Efficient LoRA
%A Wang, Sheng
%A Xue, Boyang
%A Ye, Jiacheng
%A Jiang, Jiyue
%A Chen, Liheng
%A Kong, Lingpeng
%A Wu, Chuan
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand
%F wang-etal-2024-prolora
%X With the rapid scaling of large language models (LLMs), serving numerouslow-rank adaptations (LoRAs) concurrently has become increasingly impractical,leading to unaffordable costs and necessitating more parameter-efficientfinetuning methods. In this work, we introduce \textPartially Rotation-enhanced Low-Rank Adaptation (PRoLoRA), an intra-layer sharing mechanism comprising fouressential components: broadcast reduction, rotation enhancement,partially-sharing refinement, and rectified initialization strategy. As asuperset of LoRA, PRoLoRA retains its advantages, and effectively circumventthe drawbacks of peer parameter-sharing methods with superior model capacity,practical feasibility, and broad applicability. Empirical experimentsdemonstrate the remarkably higher parameter efficiency of PRoLoRA in bothspecific parameter budget and performance target scenarios, and its scalabilityto larger LLMs. Notably, with one time less trainable parameters, PRoLoRA stilloutperforms LoRA on multiple instruction tuning datasets. Subsequently, anablation study is conducted to validate the necessity of individual componentsand highlight the superiority of PRoLoRA over three potential variants.Hopefully, the conspicuously higher parameter efficiency can establish PRoLoRAas a resource-friendly alternative to LoRA.
%U https://aclanthology.org/2024.acl-long.156
%P 2829-2841
Markdown (Informal)
[PRoLoRA: Partial Rotation Empowers More Parameter-Efficient LoRA](https://aclanthology.org/2024.acl-long.156) (Wang et al., ACL 2024)
ACL
- Sheng Wang, Boyang Xue, Jiacheng Ye, Jiyue Jiang, Liheng Chen, Lingpeng Kong, and Chuan Wu. 2024. PRoLoRA: Partial Rotation Empowers More Parameter-Efficient LoRA. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2829–2841, Bangkok, Thailand. Association for Computational Linguistics.