medical-vlm

OpenMedQ is a fully-open medical vision-language model pretrained on 14 datasets (~3.35M samples), achieving state-of-the-art results on medical VQA and classification benchmarks.