@inproceedings{sun-etal-2023-multi,
title = "A Multi-dimensional Evaluation of Tokenizer-free Multilingual Pretrained Models",
author = "Sun, Jimin and
Fernandes, Patrick and
Wang, Xinyi and
Neubig, Graham",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.128/",
doi = "10.18653/v1/2023.findings-eacl.128",
pages = "1725--1735",
abstract = "Recent works on tokenizer-free multilingual pretrained models show promising results in improving cross-lingual transfer and reducing engineering overhead compared to subword-based alternatives. However, previous work mainly focuses on reporting accuracy on a limited set of tasks and data settings, placing less emphasis on other important factors when tuning and deploying the models in practice, such as memory usage, inference speed, and finetuning data efficiency. We attempt to fill this gap by performing a comprehensive empirical comparison of multilingual tokenizer-free and subword-based models considering the various dimensions. Surprisingly, we find that subword-based models might still be the most practical choice in many settings, achieving better performance for lower inference latency and memory usage. Based on these results, we encourage future work in tokenizer-free methods to consider these factors when designing and evaluating new models."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sun-etal-2023-multi">
<titleInfo>
<title>A Multi-dimensional Evaluation of Tokenizer-free Multilingual Pretrained Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jimin</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Patrick</namePart>
<namePart type="family">Fernandes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinyi</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Graham</namePart>
<namePart type="family">Neubig</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EACL 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Andreas</namePart>
<namePart type="family">Vlachos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Isabelle</namePart>
<namePart type="family">Augenstein</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Dubrovnik, Croatia</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Recent works on tokenizer-free multilingual pretrained models show promising results in improving cross-lingual transfer and reducing engineering overhead compared to subword-based alternatives. However, previous work mainly focuses on reporting accuracy on a limited set of tasks and data settings, placing less emphasis on other important factors when tuning and deploying the models in practice, such as memory usage, inference speed, and finetuning data efficiency. We attempt to fill this gap by performing a comprehensive empirical comparison of multilingual tokenizer-free and subword-based models considering the various dimensions. Surprisingly, we find that subword-based models might still be the most practical choice in many settings, achieving better performance for lower inference latency and memory usage. Based on these results, we encourage future work in tokenizer-free methods to consider these factors when designing and evaluating new models.</abstract>
<identifier type="citekey">sun-etal-2023-multi</identifier>
<identifier type="doi">10.18653/v1/2023.findings-eacl.128</identifier>
<location>
<url>https://aclanthology.org/2023.findings-eacl.128/</url>
</location>
<part>
<date>2023-05</date>
<extent unit="page">
<start>1725</start>
<end>1735</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Multi-dimensional Evaluation of Tokenizer-free Multilingual Pretrained Models
%A Sun, Jimin
%A Fernandes, Patrick
%A Wang, Xinyi
%A Neubig, Graham
%Y Vlachos, Andreas
%Y Augenstein, Isabelle
%S Findings of the Association for Computational Linguistics: EACL 2023
%D 2023
%8 May
%I Association for Computational Linguistics
%C Dubrovnik, Croatia
%F sun-etal-2023-multi
%X Recent works on tokenizer-free multilingual pretrained models show promising results in improving cross-lingual transfer and reducing engineering overhead compared to subword-based alternatives. However, previous work mainly focuses on reporting accuracy on a limited set of tasks and data settings, placing less emphasis on other important factors when tuning and deploying the models in practice, such as memory usage, inference speed, and finetuning data efficiency. We attempt to fill this gap by performing a comprehensive empirical comparison of multilingual tokenizer-free and subword-based models considering the various dimensions. Surprisingly, we find that subword-based models might still be the most practical choice in many settings, achieving better performance for lower inference latency and memory usage. Based on these results, we encourage future work in tokenizer-free methods to consider these factors when designing and evaluating new models.
%R 10.18653/v1/2023.findings-eacl.128
%U https://aclanthology.org/2023.findings-eacl.128/
%U https://doi.org/10.18653/v1/2023.findings-eacl.128
%P 1725-1735
Markdown (Informal)
[A Multi-dimensional Evaluation of Tokenizer-free Multilingual Pretrained Models](https://aclanthology.org/2023.findings-eacl.128/) (Sun et al., Findings 2023)
ACL