boffire/kabyle-english-translatewiki
收藏资源简介:
这是一个从translatewiki.net批量转储(2026-01-01)中提取的英语到卡拜尔语(Taqbaylit)的平行语料库,经过清洗和去重处理。数据集包含8,871个独特的翻译对,主要用于软件本地化、用户界面字符串和文档翻译。语言对为英语(en)作为源语言,卡拜尔语(kab)作为目标语言,卡拜尔语是一种主要在阿尔及利亚卡比利亚地区使用的柏柏尔语(Tamazight)。数据被划分为训练集(约7,096对,占80%)、验证集(约887对,占10%)和测试集(约888对,占10%)。处理步骤包括下载、提取、解析GNU gettext .po文件、去重、过滤空字符串和拆分,最终从原始12,603对中移除3,732个重复项并解决1,020个冲突。数据集在CC BY 3.0许可证下发布。
A clean, deduplicated parallel corpus of English → Kabyle (Taqbaylit) translations extracted from the translatewiki.net bulk dump (2026-01-01). It contains 8,871 unique pairs for software localization, UI strings, and documentation. The language pair is English (en) as source and Kabyle (kab) as target, with Kabyle being a Berber (Tamazight) language spoken primarily in Kabylia, Algeria. The data is split into train (~7,096 pairs, 80%), validation (~887 pairs, 10%), and test (~888 pairs, 10%). Processing steps involved downloading, extracting, parsing GNU gettext .po files, deduplication, filtering empty strings, and splitting, resulting in removal of 3,732 duplicates and resolution of 1,020 conflicts from 12,603 raw pairs. The dataset is released under CC BY 3.0 license.



