Skip to content

Dataset · FreedomIntelligence

Large medical image and caption corpus

FreedomIntelligence/PubMedVision-Plus

Large-scale English dataset of tens of millions of medical image-caption pairs from PubMed articles.

What's inside

Contains pairs of scientific illustrations and author expert descriptions for multimodal pre-training in medicine. The volume ranges from 10 to 100 million English examples in image and text formats. Data extracted from approximately 15 million PubMed articles and further cleaned using language and vision-language models (LLMs and VLMs).

What you can do with it

  • Pre-training and fine-tuning medical vision-language models (VLMs)
  • Training caption generation models for biomedical scans and charts
  • Building search systems for scientific medical literature incorporating illustrations
Tasks
Image → text
Languages
English
Size
10M–100M rows
Downloads this month
689
Likes
2

How to load it

from datasets import load_dataset

ds = load_dataset("FreedomIntelligence/PubMedVision-Plus")