@inproceedings{liu-etal-2024-e2,
title = "E2-{LLM}: Efficient and Extreme Length Extension of Large Language Models",
author = "Liu, Jiaheng and
ZhiqiBai, ZhiqiBai and
Zhang, Yuanxing and
Zhang, Chenchen and
YuangZh, YuangZh and
Zhang, Ge and
JiakaiWang, JiakaiWang and
Que, Haoran and
Chen, Yukang and
Su, Wenbo and
Ge, Tiezheng and
Fu, Jie and
Chen, Wenhu and
Zheng, Bo",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.252",
pages = "4243--4253",
abstract = "Training Large Language Models (LLMs) to process extensive context lengths incurs prohibitive computational costs. Prevailing techniques for extending context capabilities in LLMs typically require not only additional training procedures but also access to datasets with long context (e.g., sequences of 32K tokens), presupposing substantial GPU expenditures. To address the aforementioned issues, we introduce a novel solution named Efficient and Extreme length extension for Large Language Models (E2-LLM). E2-LLM entails a singular training process over considerably short sequences (e.g., 4K tokens), which greatly mitigates the cost of continual-pretraining or fine-tuning. Within the training phase, we incorporate a dual augmentation strategy with Rotary Position Embeddings (RoPE) that adjusts the scale and position indices across distinct training samples. E 2 -LLM is meticulously designed to enhance the model{'}s robustness to diverse relative positions. The experimental results on multiple benchmark datasets demonstrate the superior performance of E 2 -LLM on demanding tasks of processing long contexts.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2024-e2">
<titleInfo>
<title>E2-LLM: Efficient and Extreme Length Extension of Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jiaheng</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">ZhiqiBai</namePart>
<namePart type="family">ZhiqiBai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuanxing</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chenchen</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">YuangZh</namePart>
<namePart type="family">YuangZh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ge</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">JiakaiWang</namePart>
<namePart type="family">JiakaiWang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoran</namePart>
<namePart type="family">Que</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yukang</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenbo</namePart>
<namePart type="family">Su</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tiezheng</namePart>
<namePart type="family">Ge</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jie</namePart>
<namePart type="family">Fu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenhu</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bo</namePart>
<namePart type="family">Zheng</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>Findings of the Association for Computational Linguistics ACL 2024</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 and virtual meeting</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Training Large Language Models (LLMs) to process extensive context lengths incurs prohibitive computational costs. Prevailing techniques for extending context capabilities in LLMs typically require not only additional training procedures but also access to datasets with long context (e.g., sequences of 32K tokens), presupposing substantial GPU expenditures. To address the aforementioned issues, we introduce a novel solution named Efficient and Extreme length extension for Large Language Models (E2-LLM). E2-LLM entails a singular training process over considerably short sequences (e.g., 4K tokens), which greatly mitigates the cost of continual-pretraining or fine-tuning. Within the training phase, we incorporate a dual augmentation strategy with Rotary Position Embeddings (RoPE) that adjusts the scale and position indices across distinct training samples. E 2 -LLM is meticulously designed to enhance the model’s robustness to diverse relative positions. The experimental results on multiple benchmark datasets demonstrate the superior performance of E 2 -LLM on demanding tasks of processing long contexts.</abstract>
<identifier type="citekey">liu-etal-2024-e2</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.252</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>4243</start>
<end>4253</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T E2-LLM: Efficient and Extreme Length Extension of Large Language Models
%A Liu, Jiaheng
%A ZhiqiBai, ZhiqiBai
%A Zhang, Yuanxing
%A Zhang, Chenchen
%A YuangZh, YuangZh
%A Zhang, Ge
%A JiakaiWang, JiakaiWang
%A Que, Haoran
%A Chen, Yukang
%A Su, Wenbo
%A Ge, Tiezheng
%A Fu, Jie
%A Chen, Wenhu
%A Zheng, Bo
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Findings of the Association for Computational Linguistics ACL 2024
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand and virtual meeting
%F liu-etal-2024-e2
%X Training Large Language Models (LLMs) to process extensive context lengths incurs prohibitive computational costs. Prevailing techniques for extending context capabilities in LLMs typically require not only additional training procedures but also access to datasets with long context (e.g., sequences of 32K tokens), presupposing substantial GPU expenditures. To address the aforementioned issues, we introduce a novel solution named Efficient and Extreme length extension for Large Language Models (E2-LLM). E2-LLM entails a singular training process over considerably short sequences (e.g., 4K tokens), which greatly mitigates the cost of continual-pretraining or fine-tuning. Within the training phase, we incorporate a dual augmentation strategy with Rotary Position Embeddings (RoPE) that adjusts the scale and position indices across distinct training samples. E 2 -LLM is meticulously designed to enhance the model’s robustness to diverse relative positions. The experimental results on multiple benchmark datasets demonstrate the superior performance of E 2 -LLM on demanding tasks of processing long contexts.
%U https://aclanthology.org/2024.findings-acl.252
%P 4243-4253
Markdown (Informal)
[E2-LLM: Efficient and Extreme Length Extension of Large Language Models](https://aclanthology.org/2024.findings-acl.252) (Liu et al., Findings 2024)
ACL
- Jiaheng Liu, ZhiqiBai ZhiqiBai, Yuanxing Zhang, Chenchen Zhang, YuangZh YuangZh, Ge Zhang, JiakaiWang JiakaiWang, Haoran Que, Yukang Chen, Wenbo Su, Tiezheng Ge, Jie Fu, Wenhu Chen, and Bo Zheng. 2024. E2-LLM: Efficient and Extreme Length Extension of Large Language Models. In Findings of the Association for Computational Linguistics ACL 2024, pages 4243–4253, Bangkok, Thailand and virtual meeting. Association for Computational Linguistics.