Text corpus Open download PubMed metadata public domain; PMC OA articles under CC licences per article

PubMed / PMC Open Access Subset

The literature corpus behind BiomedBERT, BiomedCLIP and CONCH's caption data; the entry point for any oncology NLP or vision-language pretraining.

Dataset site { } JSON

At a glance

ProviderUS National Library of Medicine

AccessOpen download

LicencePubMed metadata public domain; PMC OA articles under CC licences per article

citations (PubMed); millions of full texts in PMC OA36 000 000

Size notesabstracts for all of PubMed; full text and figures for the open-access subset

Formatstext, JSON, PNG

Data typeBiomedical literature

TaskInformation extraction Question answering Image–text retrieval

CancerPan-cancer

Labels and annotations

MeSH indexing, article metadata, figure captions; no oncology-specific labels beyond MeSH.

Details

Not yet documented on this card.

Models trained or evaluated on it

pretraining BiomedCLIP Microsoft Research
pretraining BiomedBERT (PubMedBERT) Microsoft Research
pretraining CONCH Mahmood Lab, Brigham and Women's Hospital / Harvard Medical School — figure–caption pairs from the PMC open-access subset
training CONCH Mahmood Lab, Brigham and Women's Hospital / Harvard Medical School
training BiomedCLIP Microsoft Research

Sources

PMC Open Access Subset

This page is educational — it is not medical advice and does not replace consultation with an oncologist. Diagnostic and treatment decisions are made solely by specialist physicians.