anonimoussubmission/submission_2188_jsonl

这是一个用于古埃及语的问答基准和FLAN风格训练数据集，以JSONL格式提供，包含训练集和测试集文件。数据集旨在支持训练脚本直接读取，记录数量分别为6,164,024条（训练集）和565,672条（测试集）。它涵盖自由形式问答和FLAN风格任务，部分测试集记录由埃及学家专家锚定。数据集支持英语和古埃及语，适用于低资源语言处理，标签包括古埃及语、埃及学、问答、低资源、FLAN风格等。许可证为CC BY-SA 4.0，源语料库来自Thesaurus Linguae Aegyptiae。Question-answering benchmark and FLAN-style training dataset for Ancient Egyptian, provided in JSONL format for use with training scripts. Includes train and test splits with 6,164,024 and 565,672 records respectively, covering free-form QA and FLAN-style tasks. The test set contains expert-anchor records by Egyptologists. Supports English and Ancient Egyptian languages, tagged for ancient-Egyptian, Egyptology, question-answering, low-resource, and flan-style. Licensed under CC BY-SA 4.0, with source corpus from Thesaurus Linguae Aegyptiae.