Enregistré dans:
Détails bibliographiques
Auteurs principaux: Lin, Cheng-Wei, Hsieh, Wan-Hsuan, Guan, Kai-Xin, Hsu, Chan-Jan, Kuo, Chia-Chen, Lai, Chuan-Lin, Chung, Chung-Wei, Wang, Ming-Jen, Shiu, Da-Shan
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2411.16387
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910713372475392
author Lin, Cheng-Wei
Hsieh, Wan-Hsuan
Guan, Kai-Xin
Hsu, Chan-Jan
Kuo, Chia-Chen
Lai, Chuan-Lin
Chung, Chung-Wei
Wang, Ming-Jen
Shiu, Da-Shan
author_facet Lin, Cheng-Wei
Hsieh, Wan-Hsuan
Guan, Kai-Xin
Hsu, Chan-Jan
Kuo, Chia-Chen
Lai, Chuan-Lin
Chung, Chung-Wei
Wang, Ming-Jen
Shiu, Da-Shan
contents The quality and size of a pretraining dataset significantly influence the performance of large language models (LLMs). While there have been numerous efforts in the curation of such a dataset for English users, there is a relative lack of similar initiatives for Traditional Chinese. Building upon this foundation of FineWeb, we introduce FineWeb-zhtw, a dataset tailored specifically for Traditional Chinese users. We came up with multiple stages of meticulously designed filters to cater to the linguistic difference between English and Traditional Chinese, to ensure comprehensiveness and quality. We determined effectiveness from querying dataset samples with three main objectives. Our code and datasets are publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2411_16387
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FineWeb-zhtw: Scalable Curation of Traditional Chinese Text Data from the Web
Lin, Cheng-Wei
Hsieh, Wan-Hsuan
Guan, Kai-Xin
Hsu, Chan-Jan
Kuo, Chia-Chen
Lai, Chuan-Lin
Chung, Chung-Wei
Wang, Ming-Jen
Shiu, Da-Shan
Computation and Language
Databases
The quality and size of a pretraining dataset significantly influence the performance of large language models (LLMs). While there have been numerous efforts in the curation of such a dataset for English users, there is a relative lack of similar initiatives for Traditional Chinese. Building upon this foundation of FineWeb, we introduce FineWeb-zhtw, a dataset tailored specifically for Traditional Chinese users. We came up with multiple stages of meticulously designed filters to cater to the linguistic difference between English and Traditional Chinese, to ensure comprehensiveness and quality. We determined effectiveness from querying dataset samples with three main objectives. Our code and datasets are publicly available.
title FineWeb-zhtw: Scalable Curation of Traditional Chinese Text Data from the Web
topic Computation and Language
Databases
url https://arxiv.org/abs/2411.16387