ghananlpcommunity/twi-english-parallel-text
收藏资源简介:
Twi–英语平行文本数据集是一个由Ghana NLP Community编译的Asante Twi和英语平行语料库。该数据集包含句子对,来源包括YouVersion Bible的经文级平行对(使用Asante Twi圣经版本1461和1861,与CEB英语圣经对齐)以及历史词典和数字化档案的句子对。数据集特征包括twi(Asante Twi句子)、eng(对应英语句子)、count(确认该对的独立来源数量)和source_files(源文件标识符的分号分隔列表)。它主要用于翻译任务,支持Twi(特维语)和英语,规模在10K到100K个示例之间,属于低资源语言数据集,适用于自然语言处理研究和应用。
The Twi–English Parallel Text Dataset is a parallel corpus of Asante Twi and English sentence pairs compiled by the Ghana NLP Community. It includes verse-level parallel pairs scraped from YouVersion Bible (versions 1461 & 1861 of the Asante Twi Bible, aligned with the CEB English Bible) and sentence pairs from historical dictionaries and digitised archives. The dataset features columns for twi (Asante Twi sentence), eng (corresponding English sentence), count (number of independent sources that confirmed the pair), and source_files (semicolon-separated list of source identifiers). It is designed for translation tasks, supports Twi and English, has a size category of 10K<n<100K, and is tagged for low-resource language applications.




