@inproceedings{vinagre-etal-2026-fine,
title = "Fine {PT}-{PT} Web: A High-Quality 41 Billion Tokens Data Collection of the {E}uropean {P}ortuguese Web",
author = "Vinagre, Gon{\c{c}}alo and
Guerra, Rui Pedro and
Gomes, Pedro and
Ramos, Miguel Moura and
Alves, Duarte Miguel and
Simpl{\'i}cio, Afonso and
Tavares, Diogo and
Semedo, David and
Gomes, Daniel and
Magalhaes, Joao",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2418/",
pages = "54905--54920",
ISBN = "979-8-89176-459-0",
abstract = "Curating Web corpora for regional language variants like European Portuguese (PT-PT) is heavily bottlenecked by dialectal overlap (mainly with PT-BR) and data processing scale. This paper presents an efficient pipeline to curate a production-ready PT-PT corpus from the Portuguese Web, spanning 411 TB of raw data from Arquivo.pt. We introduce a novel post-scraping block that removes boilerplate and line duplicates prior to filtering. This early-stage intervention increases final document yield by 19.04{\%} by rescuing valid text that standard heuristic filters prematurely discard. Integrated with rigorous language identification, weighted fuzzy deduplication, and neural quality classification, our pipeline offers a scalable framework and a clean, representative corpus optimized for LLM pre-training."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="vinagre-etal-2026-fine">
<titleInfo>
<title>Fine PT-PT Web: A High-Quality 41 Billion Tokens Data Collection of the European Portuguese Web</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gonçalo</namePart>
<namePart type="family">Vinagre</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="given">Pedro</namePart>
<namePart type="family">Guerra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pedro</namePart>
<namePart type="family">Gomes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miguel</namePart>
<namePart type="given">Moura</namePart>
<namePart type="family">Ramos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Duarte</namePart>
<namePart type="given">Miguel</namePart>
<namePart type="family">Alves</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Afonso</namePart>
<namePart type="family">Simplício</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Diogo</namePart>
<namePart type="family">Tavares</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Semedo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daniel</namePart>
<namePart type="family">Gomes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joao</namePart>
<namePart type="family">Magalhaes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Curating Web corpora for regional language variants like European Portuguese (PT-PT) is heavily bottlenecked by dialectal overlap (mainly with PT-BR) and data processing scale. This paper presents an efficient pipeline to curate a production-ready PT-PT corpus from the Portuguese Web, spanning 411 TB of raw data from Arquivo.pt. We introduce a novel post-scraping block that removes boilerplate and line duplicates prior to filtering. This early-stage intervention increases final document yield by 19.04% by rescuing valid text that standard heuristic filters prematurely discard. Integrated with rigorous language identification, weighted fuzzy deduplication, and neural quality classification, our pipeline offers a scalable framework and a clean, representative corpus optimized for LLM pre-training.</abstract>
<identifier type="citekey">vinagre-etal-2026-fine</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2418/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>54905</start>
<end>54920</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Fine PT-PT Web: A High-Quality 41 Billion Tokens Data Collection of the European Portuguese Web
%A Vinagre, Gonçalo
%A Guerra, Rui Pedro
%A Gomes, Pedro
%A Ramos, Miguel Moura
%A Alves, Duarte Miguel
%A Simplício, Afonso
%A Tavares, Diogo
%A Semedo, David
%A Gomes, Daniel
%A Magalhaes, Joao
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F vinagre-etal-2026-fine
%X Curating Web corpora for regional language variants like European Portuguese (PT-PT) is heavily bottlenecked by dialectal overlap (mainly with PT-BR) and data processing scale. This paper presents an efficient pipeline to curate a production-ready PT-PT corpus from the Portuguese Web, spanning 411 TB of raw data from Arquivo.pt. We introduce a novel post-scraping block that removes boilerplate and line duplicates prior to filtering. This early-stage intervention increases final document yield by 19.04% by rescuing valid text that standard heuristic filters prematurely discard. Integrated with rigorous language identification, weighted fuzzy deduplication, and neural quality classification, our pipeline offers a scalable framework and a clean, representative corpus optimized for LLM pre-training.
%U https://aclanthology.org/2026.emnlp-main.2418/
%P 54905-54920
Markdown (Informal)
[Fine PT-PT Web: A High-Quality 41 Billion Tokens Data Collection of the European Portuguese Web](https://aclanthology.org/2026.emnlp-main.2418/) (Vinagre et al., EMNLP 2026)
ACL
- Gonçalo Vinagre, Rui Pedro Guerra, Pedro Gomes, Miguel Moura Ramos, Duarte Miguel Alves, Afonso Simplício, Diogo Tavares, David Semedo, Daniel Gomes, and Joao Magalhaes. 2026. Fine PT-PT Web: A High-Quality 41 Billion Tokens Data Collection of the European Portuguese Web. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 54905–54920, Budapest, Hungary. Association for Computational Linguistics.