metadata
license: apache-2.0
task_categories:
- visual-question-answering
- image-to-text
language:
- en
tags:
- medical
- radiology
- vision-language
Medical Vision-Language Dataset
Combined dataset for training medical vision-language models.
Statistics
- Total: 3793 samples
- Train: 3035 | Validation: 758
Sources
- ROCO: 2000
- VQA-RAD: 1793
Created: 2025-11-23 22:25:36