<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2025-efficient-ensemble">
    <titleInfo>
        <title>Efficient Ensemble for Fine-tuning Language Models on Multiple Datasets</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Dongyue</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ziniu</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Lu</namePart>
        <namePart type="family">Wang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hongyang</namePart>
        <namePart type="given">R</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Wanxiang</namePart>
            <namePart type="family">Che</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Joyce</namePart>
            <namePart type="family">Nabende</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ekaterina</namePart>
            <namePart type="family">Shutova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mohammad</namePart>
            <namePart type="given">Taher</namePart>
            <namePart type="family">Pilehvar</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Vienna, Austria</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-251-0</identifier>
    </relatedItem>
    <abstract>This paper develops an ensemble method for fine-tuning a language model to multiple datasets. Existing methods, such as quantized LoRA (QLoRA), are efficient when adapting to a single dataset. When training on multiple datasets of different tasks, a common setup in practice, it remains unclear how to design an efficient adaptation for fine-tuning language models. We propose to use an ensemble of multiple smaller adapters instead of a single adapter per task. We design an efficient algorithm that partitions n datasets into m groups, where m is typically much smaller than n in practice, and train one adapter for each group before taking a weighted combination to form the ensemble. The algorithm leverages a first-order approximation property of low-rank adaptation to quickly obtain the fine-tuning performances of dataset combinations since methods like LoRA stay close to the base model. Hence, we use the gradients of the base model to estimate its behavior during fine-tuning. Empirically, this approximation holds with less than 1% error on models with up to 34 billion parameters, leading to an estimation of true fine-tuning performances under 5% error while speeding up computation compared to base fine-tuning by 105 times. When applied to fine-tune Llama and GPT models on ten text classification tasks, our approach provides up to 10% higher average test accuracy over QLoRA, with only 9% more FLOPs. On a Llama model with 34 billion parameters, an ensemble of QLoRA increases test accuracy by 3% compared to QLoRA, with only 8% more FLOPs.</abstract>
    <identifier type="citekey">li-etal-2025-efficient-ensemble</identifier>
    <identifier type="doi">10.18653/v1/2025.acl-long.1231</identifier>
    <location>
        <url>https://aclanthology.org/2025.acl-long.1231/</url>
    </location>
    <part>
        <date>2025-07</date>
        <extent unit="page">
            <start>25347</start>
            <end>25364</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Efficient Ensemble for Fine-tuning Language Models on Multiple Datasets
%A Li, Dongyue
%A Zhang, Ziniu
%A Wang, Lu
%A Zhang, Hongyang R.
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F li-etal-2025-efficient-ensemble
%X This paper develops an ensemble method for fine-tuning a language model to multiple datasets. Existing methods, such as quantized LoRA (QLoRA), are efficient when adapting to a single dataset. When training on multiple datasets of different tasks, a common setup in practice, it remains unclear how to design an efficient adaptation for fine-tuning language models. We propose to use an ensemble of multiple smaller adapters instead of a single adapter per task. We design an efficient algorithm that partitions n datasets into m groups, where m is typically much smaller than n in practice, and train one adapter for each group before taking a weighted combination to form the ensemble. The algorithm leverages a first-order approximation property of low-rank adaptation to quickly obtain the fine-tuning performances of dataset combinations since methods like LoRA stay close to the base model. Hence, we use the gradients of the base model to estimate its behavior during fine-tuning. Empirically, this approximation holds with less than 1% error on models with up to 34 billion parameters, leading to an estimation of true fine-tuning performances under 5% error while speeding up computation compared to base fine-tuning by 105 times. When applied to fine-tune Llama and GPT models on ten text classification tasks, our approach provides up to 10% higher average test accuracy over QLoRA, with only 9% more FLOPs. On a Llama model with 34 billion parameters, an ensemble of QLoRA increases test accuracy by 3% compared to QLoRA, with only 8% more FLOPs.
%R 10.18653/v1/2025.acl-long.1231
%U https://aclanthology.org/2025.acl-long.1231/
%U https://doi.org/10.18653/v1/2025.acl-long.1231
%P 25347-25364

Markdown (Informal)

[Efficient Ensemble for Fine-tuning Language Models on Multiple Datasets](https://aclanthology.org/2025.acl-long.1231/) (Li et al., ACL 2025)

• Efficient Ensemble for Fine-tuning Language Models on Multiple Datasets (Li et al., ACL 2025)

ACL

• Dongyue Li, Ziniu Zhang, Lu Wang, and Hongyang R. Zhang. 2025. Efficient Ensemble for Fine-tuning Language Models on Multiple Datasets. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 25347–25364, Vienna, Austria. Association for Computational Linguistics.