ConvDecoder#

class deepinv.models.ConvDecoder(img_size, in_size=(4, 4), layers=7, channels=256)[source]#

Bases: Module

Convolutional decoder network. Supports 2D and 3D data, depending on img_size & in_size

The architecture was introduced by Darestani and Heckel[1], and it is well suited as a deep image prior (see deepinv.models.DeepImagePrior).

Parameters:

  • img_size (tuple) – shape of the output image.

  • in_size (tuple) – size of the input vector.

  • layers (int) – number of layers in the network.

  • channels (int) – number of channels in the network.

References:

forward(x, scale_out=1)[source]#

Forward pass through the ConvDecoder network.

Parameters:

  • x (torch.Tensor) – Input tensor.

  • scale_out (float) – Output scaling factor.

Examples using ConvDecoder:#

Reconstructing an image using the deep image prior.

Reconstructing an image using the deep image prior.