@inproceedings{yano-etal-2025-step,
title = "{STEP}: Staged Parameter-Efficient Pre-training for Large Language Models",
author = "Yano, Kazuki and
Ito, Takumi and
Suzuki, Jun",
editor = "Chiruzzo, Luis and
Ritter, Alan and
Wang, Lu",
booktitle = "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)",
month = apr,
year = "2025",
address = "Albuquerque, New Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.naacl-short.32/",
doi = "10.18653/v1/2025.naacl-short.32",
pages = "374--384",
ISBN = "979-8-89176-190-2",
abstract = "Pre-training large language models (LLMs) faces significant memory challenges due to the large size of model weights. We introduce STaged parameter-Efficient Pre-training (STEP), which integrates parameter-efficient tuning techniques with model growth. We conduct experiments on pre-training LLMs of various sizes and demonstrate that STEP achieves up to a 53.9{\%} reduction in maximum memory requirements compared to vanilla pre-training while maintaining equivalent performance. Furthermore, we show that the model by STEP performs comparably to vanilla pre-trained models on downstream tasks after instruction tuning."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yano-etal-2025-step">
<titleInfo>
<title>STEP: Staged Parameter-Efficient Pre-training for Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kazuki</namePart>
<namePart type="family">Yano</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Takumi</namePart>
<namePart type="family">Ito</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Luis</namePart>
<namePart type="family">Chiruzzo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alan</namePart>
<namePart type="family">Ritter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Albuquerque, New Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-190-2</identifier>
</relatedItem>
<abstract>Pre-training large language models (LLMs) faces significant memory challenges due to the large size of model weights. We introduce STaged parameter-Efficient Pre-training (STEP), which integrates parameter-efficient tuning techniques with model growth. We conduct experiments on pre-training LLMs of various sizes and demonstrate that STEP achieves up to a 53.9% reduction in maximum memory requirements compared to vanilla pre-training while maintaining equivalent performance. Furthermore, we show that the model by STEP performs comparably to vanilla pre-trained models on downstream tasks after instruction tuning.</abstract>
<identifier type="citekey">yano-etal-2025-step</identifier>
<identifier type="doi">10.18653/v1/2025.naacl-short.32</identifier>
<location>
<url>https://aclanthology.org/2025.naacl-short.32/</url>
</location>
<part>
<date>2025-04</date>
<extent unit="page">
<start>374</start>
<end>384</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T STEP: Staged Parameter-Efficient Pre-training for Large Language Models
%A Yano, Kazuki
%A Ito, Takumi
%A Suzuki, Jun
%Y Chiruzzo, Luis
%Y Ritter, Alan
%Y Wang, Lu
%S Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-190-2
%F yano-etal-2025-step
%X Pre-training large language models (LLMs) faces significant memory challenges due to the large size of model weights. We introduce STaged parameter-Efficient Pre-training (STEP), which integrates parameter-efficient tuning techniques with model growth. We conduct experiments on pre-training LLMs of various sizes and demonstrate that STEP achieves up to a 53.9% reduction in maximum memory requirements compared to vanilla pre-training while maintaining equivalent performance. Furthermore, we show that the model by STEP performs comparably to vanilla pre-trained models on downstream tasks after instruction tuning.
%R 10.18653/v1/2025.naacl-short.32
%U https://aclanthology.org/2025.naacl-short.32/
%U https://doi.org/10.18653/v1/2025.naacl-short.32
%P 374-384
Markdown (Informal)
[STEP: Staged Parameter-Efficient Pre-training for Large Language Models](https://aclanthology.org/2025.naacl-short.32/) (Yano et al., NAACL 2025)
ACL
- Kazuki Yano, Takumi Ito, and Jun Suzuki. 2025. STEP: Staged Parameter-Efficient Pre-training for Large Language Models. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers), pages 374–384, Albuquerque, New Mexico. Association for Computational Linguistics.