ESRGANDiscriminator#

class deepinv.models.ESRGANDiscriminator(input_shape: tuple)[source]#

Bases: Module

ESRGAN Discriminator.

The ESRGAN discriminator model was originally proposed in ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (Wang et al.). Implementation taken from edongdongchen/EI.

See Imaging inverse problems with adversarial networks for how to use this for adversarial training.

Parameters:

input_shape (tuple) – shape of input image

forward(img)[source]#

Forward pass of discriminator model.

Parameters:

img (Tensor) – input image