@inproceedings{chen-etal-2024-lemon,
title = "{LEMON}: Reviving Stronger and Smaller {LM}s from Larger {LM}s with Linear Parameter Fusion",
author = "Chen, Yilong and
Shang, Junyuan and
Zhang, Zhenyu and
Cui, Shiyao and
Liu, Tingwen and
Wang, Shuohuan and
Sun, Yu and
Wu, Hua",
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.434",
pages = "8005--8019",
abstract = "In the new era of language models, small models (with billions of parameter sizes) are receiving increasing attention due to their flexibility and cost-effectiveness in deployment. However, limited by the model size, the performance of small models trained from scratch may often be unsatisfactory. Learning a stronger and smaller model with the help of larger models is an intuitive idea. Inspired by the observing modular structures in preliminary analysis, we propose LEMON to learn competent initial points for smaller models by fusing parameters from larger models, thereby laying a solid foundation for subsequent training. Specifically, the parameter fusion process involves two operators for layer and dimension, respectively, and we also introduce controllable receptive fields to model the prior parameter characteristics. In this way, the larger model could be transformed into any specific smaller scale and architecture. Starting from LLaMA 2-7B, we revive two stronger and smaller models with 1.3B and 2.7B. Experimental results demonstrate that the fusion-based method exhibits flexibility and outperforms a series of competitive baselines in terms of both effectiveness and efficiency.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2024-lemon">
<titleInfo>
<title>LEMON: Reviving Stronger and Smaller LMs from Larger LMs with Linear Parameter Fusion</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yilong</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junyuan</namePart>
<namePart type="family">Shang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenyu</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shiyao</namePart>
<namePart type="family">Cui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tingwen</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuohuan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hua</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>In the new era of language models, small models (with billions of parameter sizes) are receiving increasing attention due to their flexibility and cost-effectiveness in deployment. However, limited by the model size, the performance of small models trained from scratch may often be unsatisfactory. Learning a stronger and smaller model with the help of larger models is an intuitive idea. Inspired by the observing modular structures in preliminary analysis, we propose LEMON to learn competent initial points for smaller models by fusing parameters from larger models, thereby laying a solid foundation for subsequent training. Specifically, the parameter fusion process involves two operators for layer and dimension, respectively, and we also introduce controllable receptive fields to model the prior parameter characteristics. In this way, the larger model could be transformed into any specific smaller scale and architecture. Starting from LLaMA 2-7B, we revive two stronger and smaller models with 1.3B and 2.7B. Experimental results demonstrate that the fusion-based method exhibits flexibility and outperforms a series of competitive baselines in terms of both effectiveness and efficiency.</abstract>
<identifier type="citekey">chen-etal-2024-lemon</identifier>
<location>
<url>https://aclanthology.org/2024.acl-long.434</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>8005</start>
<end>8019</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LEMON: Reviving Stronger and Smaller LMs from Larger LMs with Linear Parameter Fusion
%A Chen, Yilong
%A Shang, Junyuan
%A Zhang, Zhenyu
%A Cui, Shiyao
%A Liu, Tingwen
%A Wang, Shuohuan
%A Sun, Yu
%A Wu, Hua
%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 chen-etal-2024-lemon
%X In the new era of language models, small models (with billions of parameter sizes) are receiving increasing attention due to their flexibility and cost-effectiveness in deployment. However, limited by the model size, the performance of small models trained from scratch may often be unsatisfactory. Learning a stronger and smaller model with the help of larger models is an intuitive idea. Inspired by the observing modular structures in preliminary analysis, we propose LEMON to learn competent initial points for smaller models by fusing parameters from larger models, thereby laying a solid foundation for subsequent training. Specifically, the parameter fusion process involves two operators for layer and dimension, respectively, and we also introduce controllable receptive fields to model the prior parameter characteristics. In this way, the larger model could be transformed into any specific smaller scale and architecture. Starting from LLaMA 2-7B, we revive two stronger and smaller models with 1.3B and 2.7B. Experimental results demonstrate that the fusion-based method exhibits flexibility and outperforms a series of competitive baselines in terms of both effectiveness and efficiency.
%U https://aclanthology.org/2024.acl-long.434
%P 8005-8019
Markdown (Informal)
[LEMON: Reviving Stronger and Smaller LMs from Larger LMs with Linear Parameter Fusion](https://aclanthology.org/2024.acl-long.434) (Chen et al., ACL 2024)
ACL
- Yilong Chen, Junyuan Shang, Zhenyu Zhang, Shiyao Cui, Tingwen Liu, Shuohuan Wang, Yu Sun, and Hua Wu. 2024. LEMON: Reviving Stronger and Smaller LMs from Larger LMs with Linear Parameter Fusion. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 8005–8019, Bangkok, Thailand. Association for Computational Linguistics.