imported>Zyksnowy：/* top */

2026-03-22T01:35:51Z

top

新页面

'''BookCorpus''' （也称'''Toronto Book Corpus''' ）是一个数据集，包含互联网上[[网页抓取|收集的]]约 11,000 本未出版书籍的文本。它是用于训练[[OpenAI]]的[[OpenAI|GPT]]初始版本的主要语料库<ref>{{Cite journal |date=2023-03-22 |title=BookCorpus |url=https://en.wikipedia.org/w/index.php?title=BookCorpus&oldid=1146107360 |journal=Wikipedia |language=en |access-date=2023-03-25 |archive-date=2023-03-25 |archive-url=https://web.archive.org/web/20230325124835/https://en.wikipedia.org/w/index.php?title=BookCorpus&oldid=1146107360 |dead-url=no }}</ref>，并用作包括谷歌的[[BERT]]<ref>{{Cite journal |last=Devlin |first=Jacob |last2=Chang |first2=Ming-Wei |last3=Lee |first3=Kenton |last4=Toutanova |first4=Kristina |date=2019-05-24 |title=BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding |url=http://arxiv.org/abs/1810.04805 |journal=arXiv:1810.04805 [cs] |access-date=2023-03-25 |archive-date=2023-07-09 |archive-url=https://web.archive.org/web/20230709220301/https://arxiv.org/abs/1810.04805 |dead-url=no }}</ref>在内的，其他早期大型语言模型的训练数据。该数据集包含约 9.85 亿单词，涵盖了多种类型的书籍，包括浪漫小说、科幻小说和奇幻小说。

[[多倫多大學|多伦多大学]]和[[麻省理工学院|麻省理工]]学院的研究人员在 2015 年的一篇题为“Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books”的论文中介绍了该语料库。作者将其描述为“由尚未出版的作者撰写的免费书籍”<ref>{{Cite journal |last=Zhu |first=Yukun |last2=Kiros |first2=Ryan |last3=Zemel |first3=Rich |last4=Salakhutdinov |first4=Ruslan |last5=Urtasun |first5=Raquel |last6=Torralba |first6=Antonio |last7=Fidler |first7=Sanja |date=2015 |title=Aligning Books and Movies: Towards Story-Like Visual Explanations by Watching Movies and Reading Books |url=https://www.cv-foundation.org/openaccess/content_iccv_2015/html/Zhu_Aligning_Books_and_ICCV_2015_paper.html |access-date=2023-03-25 |archive-date=2023-02-05 |archive-url=https://web.archive.org/web/20230205222219/https://www.cv-foundation.org/openaccess/content_iccv_2015/html/Zhu_Aligning_Books_and_ICCV_2015_paper.html |dead-url=no }}</ref><ref>{{Cite news|title=Google swallows 11,000 novels to improve AI's conversation|url=https://www.theguardian.com/books/2016/sep/28/google-swallows-11000-novels-to-improve-ais-conversation|newspaper=The Guardian|date=2016-09-28|accessdate=2023-03-25|issn=0261-3077|language=en-GB|first=Richard|last=Lea|archive-date=2023-03-25|archive-url=https://web.archive.org/web/20230325130418/https://www.theguardian.com/books/2016/sep/28/google-swallows-11000-novels-to-improve-ais-conversation|dead-url=no}}</ref>。该数据集起初托管在多伦多大学的网页上。原始数据集不再公开，但多伦多大学创建了一个镜像版本 BookCorpusOpen。虽然在 2015 年的原始论文中没有记录这个网站，但现在已知的删除语料库书籍的网站是{{维基数据链接|Q7544282}}。

== 参考 ==
{{reflist}}
[[Category:機器學習資料集]]

BookCorpus - 版本历史

imported>Zyksnowy：​/* top */

imported>Zyksnowy：/* top */