r"""GAN loss computation. Args: dis_output (tensor): Discriminator outputs. t_real (bool): If ``True``, uses the real

r"""Create the logging directory Args: logdir (str): Log directory name

r"""master-only print

Compute the msid descriptor for a single sample x Arguments: x: x samples ts: temperature values k: number of neighbo

r"""One residual block layer in the main branch. Args: x (4D tensor) : Current feature map. encoded_label (list of te

r"""Do any data pre-processing here. Args: data (dict): Data used for the current iteration.

r"""Remove other people in the label map except for the current target by looking at the id in the densemask map. Args: labels (NxCxH

r"""Reset the meter values

r"""Reset the trainer (for inference) at the beginning of a sequence.

r"""Reshape input to be (B x C) X H X W. Args: x (tensor or list of tensors): Input features. Returns: x (ten

r"""Save the output images to path. Note when the generate_raw_output is FALSE. Then, first_net_G_output['fake_raw_images'] is None an

r"""Select the object/person in the dict according to the object index. Currently it's used to select the target person in OpenPose dict. Arg

r"""Set pixels of the image to the given color. Args: im (HxWxC numpy array): Canvas to draw. xx (1D numpy array): x coordinates

r"""Set sample class idx. This is not used in this class... Args: class_idx (int): Which class idx to sample from.

r"""Slice all tensors from start to end. Args: data (dict, list or tensor): Input data.

r"""Things to do before an iteration. Args: data (dict): Data used for the current iteration. current_iteration (int)

r"""Things to do before an iteration. Args: data (dict): Data used for the current iteration. current_iteration (int)

r"""Update the moving average.

r"""Wrap the networks and the optimizers with AMP DDP and (optionally) model average. Args: cfg (obj): Global configuration.

r"""Compute metrics and save them to tensorboard

r"""Write metrics to the tensorboard.

r"""Using Multi-SPADE to combine raw synthesized image with warped images. Args: encoded_label (list of tensors): Origina

r"""COCO-FUNIT Generator constructor. Args: gen_cfg (obj): Generator definition part of the yaml config file. data_cf

Print all variables.

r"""Get alexnet layers

Fast CUDA implementation of `bias_act()` using custom ops.

r"""Builds sequence of augmentation ops. Returns: (list of alb.ops): List of augmentation ops.

Return Laplacian from data or load preconstructed from path Arguments: data: samples k: number of neighbours for graph const

r"""Builds sequence of augmentation ops that will be applied to each frame in the video independently. Returns: (list of alb.ops):

r"""Choose key to replace with 'image' for input to albumentations. Returns: key (str): Chosen key to be replace with 'image'

r"""Choose key to replace with 'keypoints' for input to albumentations. Returns: key (str): Chosen key to be replace with 'keypoin

r"""Collate separated images back into sequence, grouped by keys. Args: augmented (dict): Dict containing frames with keys of the

r"""Computes required statistics about dataset. All children need to implement their own.

r"""Compute FID. We will compute a FID value per test class. That is if you have 30 test classes, we will compute 30 different FID values.

r"""Compute FID for both domains.

r"""FID computation function to be overloaded.

r"""Compute FID for both domains.

r"""Create additional targets as required by the albumentation library. Args: inputs (dict): Keys are from self.augmentable_data_

r"""Creates mapping from idx to key in LMDB. Returns: (tuple): - self.mapping (dict): Dict of seq_len to list of se

r"""Creates mapping from data sample idx to actual LMDB keys. All children need to implement their own. Returns: self

r"""Creates mapping from idx to key in LMDB. Returns: (tuple): - self.mapping (dict): Dict of seq_len to list of se

r"""Create the LMDB key for this piece of information. Args: sequence_name (str): Which sequence from the chosen dataset.

r"""All other custom losses are defined here.

r"""Make input one-hot. Args: label_map (torch.Tensor): (C, H, W) tensor containing indices. num_classes (int): Number of labels

r"""Operations to do after an epoch. Args: data (dict): Data used for the current iteration. current_epoch (int): Cur

r"""Operations to do after an iteration. Args: data (dict): Data used for the current iteration. current_epoch (int):

r"""Forwarding the MLP. Args: blk_feats (K x C1 tensor): Sparse block features. worldcoord2 (N x H x W x L x 3 tensor

r"""Sample camera poses and perform ray-voxel intersection. Args: batch_size (int): Expected batch size of the current batch

r"""All other custom losses go here. Args: net_D_output (dict): Output of the discriminator.

r"""All other custom generator losses go here. Args: data_t (dict): Training data at the current time t. net_G_output

r"""Find key-value inputs to function from string definition. Args: function_string (str): String with function name and args. e.

r""" Computes precision, recall, density, and coverage given two manifolds. Args: real_features: numpy.ndarray([N, feature_dim], dtyp

r"""Get height and width to resize to, given smallest side. Args: height (int): Input image height. width (int): Inpu

r"""Return dataset object for the test set Args: cfg (obj): Global configuration file. Returns: (obj): PyTorch dataset objec

r"""Return dataset objects for the training and validation sets. Args: cfg (obj): Global configuration file. Returns: (dict)

r"""Compute visualization outputs. Args: data (dict): Data used for the current iteration.

r"""Gets selected files. Args: index (int): Index into dataset. concat (bool): Concatenate all items in labels?

r"""Gets selected files. Args: index (int): Index into dataset. concat (bool): Concatenate all items in labels?

r"""Get inception v3 layers

r"""Initialize a dictionary of hyperparameters that we want to monitor in the HParams dashboard in tensorBoard.

r"""Every trainer should implement its own init loss function.

r"""Initialize the tensorboard. Different algorithms might require different performance metrics. Hence, custom tensorboard initializa

r"""Based on https://github.com/dougalsutherland/opt-mmd/blob/master/two_sample/mmd.py but changed to not compute the full kernel matrix at on

r"""Perform different data augmentation on different image inputs. Note that this operation only works Args: inputs (dict): Keys

r"""Computes MMD between two sets of features using polynomial kernels. It performs a number of repetitions of subset sampling without replacement

r"""Implement the things you want to do before saving a checkpoint. For example, you can compute the K-mean features (pix2pixHD) before

r"""Resize input label maps and images so that it can be properly generated by the generator. Args: data (dict): Input di

r"""Get resnet50 layers

r"""Get robust resnet50 layers

r"""Gets files to load for this sample. Args: index (int): Index in [0, len(dataset)]. Returns: key (dict):

r"""Gets files to load for this sample. Args: index (int): Index in [0, len(dataset)]. Returns: keys (dict):

r"""Gets files to load for this sample. Args: index (int): Index in [0, len(dataset)]. Returns: (tuple):

r"""Gets files to load for this sample. Args: index (int): Index in [0, len(dataset)]. Returns: key (dict):