Dict as attribute trick.

r"""A Wrapper class that wraps ``torch.nn.Conv2d`` with normalization and nonlinearity. Args: in_channels (int): Number of channels i

r"""A Wrapper class that wraps ``torch.nn.Linear`` with normalization and nonlinearity. Args: in_features (int): Number of channels i

r"""Linear layer with affine modulation (Based on StyleGAN2 mod demod). Equivalent to affine modulation following linear, but faster when the same

r"""The network that extracts features to compute the perceptual loss. Args: network (nn.Sequential) : The network that extracts features

r"""Configuration class. This should include every human specifiable hyperparameter values for your training.

r"""Add Gaussian noise to the input tensor.

Class to colorize segmentation maps.

Meter is to keep track of statistics along steps. Meters write values for purpose like printing average values. Meters can be flushed to log f

r"""Learning affine modulation of activation. Args: in_features (int): Number of input features. style_features (int): Number of

r"""GAN loss constructor. Args: gan_mode (str): Type of GAN loss. ``'hinge'``, ``'least_square'``, ``'non_saturated'``, ``'wa

r"""In this model average implementation, the spectral layers are absorbed in the model parameter by default. If such options are turned on, b

r"""SPADE Image Generator constructor. Args: num_labels (int): Number of different labels. out_image_small_side_size (int): min(w

r"""Dummy wrapping the module.

r"""Adaptive normalization layer. The layer first normalizes the input, then performs an affine transformation using parameters computed from the

r"""Compute feature matching loss

r"""Fromage optimizer implementation (https://arxiv.org/abs/2002.03432)

r"""Manages the SPADE/GauGAN model used to generate pseudo-GTs for training GANcraft. Args: gaugan_cfg (Config): SPADE configuration.

r"""Compute KL loss in VAE for Gaussian distributions

r"""Hyper Conv2d initialization. Args: in_channels (int): Dummy parameter. out_channels (int): Dummy parameter. kernel_si

r"""Spatially Adaptive Normalization (SPADE) initialization. Args: num_features (int) : Number of channels in the input tensor. c

r"""Layer Normalization as introduced in https://arxiv.org/abs/1607.06450. This is the usual way to apply layer normalization in CNNs. Not

r""" MLP with affine modulation.

r"""MADAM optimizer implementation (https://arxiv.org/abs/2006.14560)

r"""PCG Datasets

r"""Perceptual loss initialization. Args: network (str) : The name of the loss network: 'vgg16' | 'vgg19'. layers (str or list of s

r"""CNN converting intermediate feature map to final image.

r"""MLP converting ray directions to sky features.

r"""Scale normalization: "Transformers without Tears: Improving the Normalization of Self-Attention" Modified from: https://github.com/tnq

r"""Spatially Adaptive Normalization (SPADE) initialization. Args: num_features (int) : Number of channels in the input tensor. c

r"""Style Encode constructor. Args: style_enc_cfg (obj): Style encoder definition file.

r"""Style Encoder constructor. Args: style_enc_cfg (obj): Style encoder definition file.

r"""Weight demodulation in "Analyzing and Improving the Image Quality of StyleGAN", Karras et al. Args: conv (torch.nn.Modules): Conv

r"""Dummy wrapping the module.

r"""Device used for nvml.

r"""Handles data augmentation using albumentations library.

r"""Minecraft 3D generator constructor. Args: gen_cfg (obj): Generator definition part of the yaml config file. data_cfg (obj): D

r"""Base trainer. We expect that all trainers inherit this class. Args: cfg (obj): Global configuration. net_G (obj): Generator n

r"""A Wrapper class that wraps ``torch.nn.Conv1d`` with normalization and nonlinearity. Args: in_channels (int): Number of channels i

r"""A Wrapper class that wraps ``torch.nn.Conv3d`` with normalization and nonlinearity. Args: in_channels (int): Number of channels i

r"""Residual block for 2D input. Args: in_channels (int) : Number of channels in the input tensor. out_channels (int) : Number of

r"""Multi-resolution patch discriminator. Based on FPSE discriminator but with N+1 labels. Args: dis_cfg (obj): Discriminator definition

r"""Residual block for 2D input with downsampling. Args: in_channels (int) : Number of channels in the input tensor. out_channels

r"""SceneDreamer generator constructor. Args: gen_cfg (obj): Generator definition part of the yaml config file. data_cfg (obj): D

r"""SPADE generator constructor. Args: gen_cfg (obj): Generator definition part of the yaml config file. data_cfg (obj): Data def

r"""A Wrapper class that wraps ``HyperConv2d`` with normalization and nonlinearity. Args: in_channels (int): Number of channels in th

r"""Hyper residual block for 2D input. Args: in_channels (int) : Number of channels in the input tensor. out_channels (int) : Num

r"""Resolving mapping across Minecraft voxel, coco-stuff label and reduced label set.

r"""A Wrapper class that wraps ``torch.nn.Conv2d`` with normalization and nonlinearity. It can return multiple outputs, if some layers in the bloc

r"""Residual block for 2D input. It can return multiple outputs, if some layers in the block return more than one output. Args: in_ch

r"""Self attention Layer Args: in_channels (int): Number of channels in the input tensor. scale (bool, optional, default=True): I

r"""Partial 2D convolution in "Image inpainting for irregular holes using partial convolutions." Liu et al., ECCV 2018

r"""A Wrapper class that wraps ``PartialConv2d`` with normalization and nonlinearity. Args: in_channels (int): Number of channels in

r"""Partial 3D convolution in "Image inpainting for irregular holes using partial convolutions." Liu et al., ECCV 2018

r"""A Wrapper class that wraps ``PartialConv3d`` with normalization and nonlinearity. Args: in_channels (int): Number of channels in

r"""Residual block for 2D input with partial convolution. Args: in_channels (int) : Number of channels in the input tensor. out_c

r"""Residual block for 3D input with partial convolution. Args: in_channels (int) : Number of channels in the input tensor. out_c

r"""Sequential block for partial convolutions.

r""" MLP with affine modulation.

r"""Residual block for 1D input. Args: in_channels (int) : Number of channels in the input tensor. out_channels (int) : Number of

r"""Residual block for 2D input. Args: in_channels (int) : Number of channels in the input tensor. out_channels (int) : Number of

r"""Residual block for 3D input. Args: in_channels (int) : Number of channels in the input tensor. out_channels (int) : Number of

r"""Residual block with full-connected layers. Args: in_channels (int) : Number of channels in the input tensor. out_channels (in

r"""MLP converting style code to intermediate style representation.

r"""Initialize GANcraft trainer. Args: cfg (Config): Global configuration. net_G (obj): Generator network. net_D (obj): D

r"""Residual block for 2D input with downsampling. Args: in_channels (int) : Number of channels in the input tensor. out_channels

r"""An abstract wrapper class that wraps a torch convolution or linear layer with normalization and nonlinearity.

r"""Compute Weighted MSE loss

r"""An abstract wrapper class that wraps a torch convolution or linear layer with normalization and nonlinearity.

r"""An abstract class for residual blocks with downsampling.