UNet#

class composer.models.UNet(num_classes=3)[source]#

A U-Net model extending ComposerModel.

See U-Net: Convolutional Networks for Biomedical Image Segmentation (Ronneberger et al, 2015) on the U-Net architecture.

Parameters

num_classes (int, optional) โ€“ The number of classes. Needed for classification tasks. Default: 3.

inference2d(image)[source]#

Runs inference on a 3D image, by passing each depth slice through the model.