Models Supported: UNet, UNet-Ensembled, UNet+, UNet++, MultiResUNet (with optional Deep Supervision and Autoencoder modes for 1D or 2D).