This is a new DistributedSampler that use more sophisticate way for random sync. Arguments: dataset: Dataset used for sampli

Modified from torch.utils.data._utils.collate It handle list different from the default. List collate just by append e

The lrelu layer with alpha, gain and clamp

This is the base functor for all types of executions. One execution can have multiple stages, but are only allowed to use the sa

Apply gaussian smoothing on a 1d, 2d or 3d tensor. Filtering is performed seperately for each channel in the input using a depthwise conv

This formatter perform the following operation. a) Random scale the image h w so it is in a designated range = [1, 1.2] or larger. i.e

This formatter that use the same center mask

This formatter that use the same center mask

r"""Collects the scalars broadcasted by `report()` and `report0()` and computes their long-term averages (mean and standard deviation) over us

This formatter is a direct replication of the original CoModGan TF code

This formatter is a direct replication of the original CoModGan TF code

Convenience class that behaves like a dict but allows access with the attribute syntax.

Loader with resolution 512

This formatter is a direct replication of the original CoModGan TF code

This formatter perform the following operation. a) Rescale the image h w to the designated resolution.

Redirect stderr to stdout, optionally print stdout to a file, and optionally force flushing on both stdout and the file.

Loader with resolution 512

This formatter is a direct replication of the original CoModGan TF code