NYTK/HuBoolQ
收藏资源简介:
BoolQ匈牙利语(机器翻译)数据集是BoolQ阅读理解数据集的匈牙利语机器翻译版本,用于是/否问题。每个示例包含原始英语问题和段落、匈牙利语翻译以及可选的修正版本,因此可用作并行语料库或根据任务需求选择合适形式。数据集结构包括原始英语问题(question_en)、匈牙利语机器翻译问题(question)、修正后匈牙利语问题(corrected_question)、答案(answer)、原始英语段落(passage_en)、匈牙利语机器翻译段落(passage)、修正后匈牙利语段落(corrected_passage)和翻译工具标识(translation_tool)。数据集包含9,427个训练示例和3,270个验证示例。构建过程包括使用Google Translate进行翻译、规则清理、拼写检查、GPT-4o mini重写和手动验证。数据集仍处于开发中,未来会进一步修正。
BoolQ Hungarian (Machine Translation) is a Hungarian translation of the BoolQ reading-comprehension dataset for yes/no questions. Every example contains the original English question and passage alongside the Hungarian translation and an optional corrected version, so you can use it as a parallel corpus or pick whichever form fits your task. The dataset structure includes original English question (question_en), Hungarian machine translation of the question (question), polished Hungarian version of the question (corrected_question), ground-truth answer (answer), original English passage (passage_en), Hungarian machine translation of the passage (passage), polished Hungarian version of the passage (corrected_passage), and translation tool identifier (translation_tool). It contains 9,427 training examples and 3,270 validation examples. The dataset was built through translation using Google Translate, rule-based cleanup, spell-checking, GPT-4o mini rewriting, and manual verification. It is still a work in progress with further corrections to be applied in the future.




