Randomly translate image horizontally with an offset sampled in [-x, x] and vertically [-y, y]. Note that the coordinate are not pixe

Unfreeze all parameters of `net`

Set weight norm and equalized learning rate like demodulated convs in StyleGAN2 for module m. The weight matrix is initialized for a lea

R1 regularizer from Which Training Methods for GANs do actually Converge? ( https://arxiv.org/abs/1801.04406 ). It forces grad norm of 0

R2 regularizer from Which Training Methods for GANs do actually Converge? ( https://arxiv.org/abs/1801.04406 ). It forces grad norm of 0

R2 regularizer from The GAN is dead; long live the GAN! A Modern Baseline GAN ( https://openreview.net/pdf?id=OrtN9hPP7V ). It forces gr

Correlate the color of the image `img` and return the result

Args: x (PIL.Image): the image to transform Returns: resized image

Pad the image Args: img (PIL.Image): The image to pad Returns: Padded image

Transform the image Args: x: image Returns: A tuple `out` so that `out[i] = transforms[i](x)`

Args: img (PIL Image): Image to be cropped and resized. Returns: PIL Image: Randomly cropped and resized imag

Detect edges Args: img (PIL.Image): the image Returns: edges detected in `img` as PIL Image

Args: i (int): index Returns: A tuple (i, self.ds[i])

Returns: The ith element of the underlying dataset or its cached value if available

Return the ith paired image as (img1, img2)