MedMNIST v2

Medical image dataset. MedMNIST v2 is a large-scale 2D and 3D medical image classification dataset, containing 12 2D datasets and 6 3D datasets, among which the 2D datasets have 708,069 images and the 3D datasets have 10,214 images. The dataset includes multiple modalities (X-ray, retinal OCT, ultrasound, CT, etc.), multiple tasks (multi-class classification, binary classification, multi-label, ordinal regression), and the dataset sizes range from hundreds to hundreds of thousands;