This repository implements a collection of 3D deep learning architectures for medical image segmentation, including U-Net, ResNet, and DenseNet variants. It provides data loaders for popular medical imaging challenges such as BraTS 2018/2019, iSeg, and MRBrains18. The library focuses on multi-modal 3D brain MRI segmentation but supports extensible architecture additions. Users can train, evaluate, and perform inference with pre-built segmentation models on medical imaging data.