PMC Large is a cleaned text dataset derived from PMC Open Access articles with permissive licenses.
Each row contains a PMC identifier, cleaned abstract, cleaned article body text, and metadata fields including PMID, DOI, article title, journal title, electronic publication date, authors, article type, and license kind.
The metadata.epub field is stored as a string because source values can be full dates such as 2003-08-18, year-month values, year-only values such as 2006… See the full description on the dataset page:
https://huggingface.co/datasets/nefelipar/pmc-large.
