lmqg/qg_itquad
收藏资源简介:
该数据集是SQuAD-it的一个修改版本,专门用于问题生成任务。由于原始数据集仅包含训练集和验证集,因此从训练集中手动抽取了测试集,以确保测试集与训练集在段落层面没有重叠。数据集的语言为意大利语,任务类型为文本生成,具体是问题生成。数据集的字段包括问题、段落、答案、句子等,并且有特殊的标记用于突出显示答案或包含答案的句子。
This dataset is a modified version of SQuAD-it, specifically designed for the question generation task. Since the original SQuAD-it dataset only contains training and validation splits, a test split was manually extracted from the original training split to ensure no paragraph-level overlap exists between the test split and the training split. The dataset is in Italian, and its task category is text generation, specifically question generation. Its dataset fields include questions, paragraphs, answers, sentences and other relevant entries, with special tokens used to highlight either the target answer or the sentence containing the answer.
数据集概述
基本信息
- 数据集名称: SQuAD-it for question generation
- 许可证: cc-by-4.0
- 语言: 意大利语 (it)
- 多语言性: 单语种
- 大小: 10K<n<100K
- 源数据集: squad_es
- 任务类别: text-generation
- 任务ID: language-modeling
- 标签: question-generation
数据集描述
- 概述: 本数据集是QG-Bench的一个子集,专为问句生成任务设计。它是SQuAD-it的一个修改版本,特别从训练集中手动抽样出一个测试集,确保与训练集在段落层面无重叠。
- 支持的任务: 问句生成 (
question-generation) - 评估指标: BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore
数据集结构
- 数据字段:
question: 字符串类型paragraph: 字符串类型answer: 字符串类型sentence: 字符串类型paragraph_answer: 字符串类型,答案通过特殊标记<hl>高亮paragraph_sentence: 字符串类型,包含答案的句子通过特殊标记<hl>高亮sentence_answer: 字符串类型,答案通过特殊标记<hl>高亮
- 数据分割:
train: 46550条记录validation: 7609条记录test: 7609条记录
引用信息
@inproceedings{ushio-etal-2022-generative, title = "{G}enerative {L}anguage {M}odels for {P}aragraph-{L}evel {Q}uestion {G}eneration", author = "Ushio, Asahi and Alva-Manchego, Fernando and Camacho-Collados, Jose", booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing", month = dec, year = "2022", address = "Abu Dhabi, U.A.E.", publisher = "Association for Computational Linguistics", }




