pubmed

PubMed汇集了超过3600万条生物医学文献引文，来源于MEDLINE、生命科学期刊和在线书籍。引文可能包含指向PubMed Central和出版商网站全文内容的链接。该资源提供英文数据，主要用于文本生成、掩码填充和文本分类等任务，也支持语言建模、掩码语言建模、文本评分和主题分类。PubMed以XML格式提供年度基线数据集，并每日更新，用户可以免费下载和使用这些数据，但需遵守美国国家医学图书馆（NLM）的使用条款，包括明确标明数据来源、不使用PubMed的商标和标识等。

PubMed curates over 36 million biomedical literature citations sourced from MEDLINE, life science journals, and online books. Citations may include links to full-text content hosted on PubMed Central and publisher websites. This resource offers English-language data, which is primarily applied to tasks such as text generation, mask filling, and text classification, and also supports language modeling, masked language modeling, text scoring, and topic classification. PubMed releases annual baseline datasets in XML format and undergoes daily updates. Users may freely download and utilize these datasets, but must comply with the terms of use stipulated by the United States National Library of Medicine (NLM), including clearly attributing the data source, refraining from using PubMed's trademarks and logos, and other relevant requirements.