odeon.nn.unet.UNetResNet

class odeon.nn.unet.UNetResNet(*args: Any, **kwargs: Any)
__init__(encoder_depth, n_classes, n_channels, num_filters=32, dropout_2d=0.2, pretrained=False, is_deconv=False)

U-Net model using ResNet(18, 34, 50, 101 or 152) encoder. UNet: https://arxiv.org/abs/1505.04597 ResNet: https://arxiv.org/abs/1512.03385 Proposed by Alexander Buslaev: https://www.linkedin.com/in/al-buslaev/

Parameters

  • encoder_depth (int) – depth of a ResNet encoder (18, 34, 50, 101 or 152).

  • n_channels (int) – number of input channels

  • n_classes (int) – number of output classes

  • num_filters (int, optional) – Number of filters in the last layer of decoder, by default 32

  • dropout_2d (float, optional) – probability factor of dropout layer before output layer, by default 0.2

  • pretrained (bool, optional) – False: no pre-trained weights are being used. True: ResNet encoder is pre-trained on ImageNet, by default False

  • is_deconv (bool, optional) – False: bilinear interpolation is used in decoder. True: deconvolution is used in decoder, by default False

Raises

NotImplementedError – [description]

Methods

__init__(encoder_depth, n_classes, n_channels)

U-Net model using ResNet(18, 34, 50, 101 or 152) encoder.

forward(x)