This project simply implements three common deep learning models (VGG16, ResNet18, ViT) for medical image classification tasks using the PathMNIST dataset.