A conv block that bundles conv/norm/activation layers. This block simplifies the usage of convolution layers, which are commonly used with a

Compose a data pipeline with a sequence of transforms. Args: transforms (list[dict | callable]): Either config dicts of transfo

ResLayer to build ResNet style backbone. Args: block (nn.Module): Residual block used to build ResLayer. num_blocks (int): Number

ResNet backbone. Please refer to the `paper <https://arxiv.org/abs/1512.03385>`__ for details. Args: depth (int): Network depth,

ShuffleNetV1 backbone. Args: groups (int, optional): The number of groups to be used in grouped 1x1 convolutions in each Shuf

Top-down heatmap simple head. paper ref: Bin Xiao et al. ``Simple Baselines for Human Pose Estimation and Tracking``. TopdownHeatmapSimpleHea

ViPNAS heatmap simple head. ViPNAS: Efficient Video Pose Estimation via Neural Architecture Search. More details can be found in the `paper

MobileNetV2 backbone. Args: widen_factor (float): Width multiplier, multiply number of channels in each layer by this amount.

MobileNetV3 backbone. Args: arch (str): Architecture of mobilnetv3, from {small, big}. Default: small. conv_cfg (dict

SEResNet backbone. Please refer to the `paper <https://arxiv.org/abs/1709.01507>`__ for details. Args: depth (int): Network dept

ShuffleNetV2 backbone. Args: widen_factor (float): Width multiplier - adjusts the number of channels in each layer by this am

Associative embedding with higher resolution head. paper ref: Bowen Cheng et al. "HigherHRNet: Scale-Aware Representation Learning for Bottom-

Top-down heatmap multi-stage head. TopdownHeatmapMultiStageHead is consisted of multiple branches, each of which has num_deconv_layers(>=0) n

VGG backbone. Args: depth (int): Depth of vgg, from {11, 13, 16, 19}. with_norm (bool): Use BatchNorm or not. num_classes

ViPNAS_ResNet backbone. "ViPNAS: Efficient Video Pose Estimation via Neural Architecture Search" More details can be found in the `paper

Registry class

Associative embedding simple head. paper ref: Alejandro Newell et al. "Associative Embedding: End-to-end Learning for Joint Detection and

InvertedResidual block for MobileNetV2. Args: in_channels (int): The input channels of the InvertedResidual block. out_channels (

Inverted Residual Block. Args: in_channels (int): The input channels of this Module. out_channels (int): The output channels of t

Bottleneck block for ResNet. Args: in_channels (int): Input channels of this block. out_channels (int): Output channels of this b

SCNet backbone. Improving Convolutional Networks with Self-Calibrated Convolutions, Jiang-Jiang Liu, Qibin Hou, Ming-Ming Cheng, Changhu Wang

Bottleneck block for ViPNAS_ResNet. Args: in_channels (int): Input channels of this block. out_channels (int): Output channels of

Basic block for VideoPose3D. Args: in_channels (int): Input channels of this block. out_channels (int): Output channels of this b

Data augmentation with random scaling & rotating. Args: rot_factor (int): Rotating to [-rotation_factor, rotation_factor] scale_f

A residual 3D convolutional block. Args: in_channels (int): Input channels of this block. out_channels (int): Output channels of

ViPNAS_MobileNetV3 backbone. "ViPNAS: Efficient Video Pose Estimation via Neural Architecture Search" More details can be found in the `paper

High-Resolution Module for HRFormer. Args: num_branches (int): The number of branches in the HRFormerModule. block (nn.Module): T

TCN backbone. Temporal Convolutional Networks. More details can be found in the `paper <https://arxiv.org/abs/1811.11742>`__ . Args:

BasicBlock for ResNet. Args: in_channels (int): Input channels of this block. out_channels (int): Output channels of this block.

A conv block that bundles conv/norm/activation layers. This block simplifies the usage of convolution layers, which are commonly used with a

Code modified from https://github.com/shubham-goel/4D-Humans/blob/a0def798c7eac811a63c8220fcc22d983b39785e/hmr2/models/components/pose_transforme

RegNet backbone. More details can be found in `paper <https://arxiv.org/abs/2003.13678>`__ . Args: arch (dict): The parameter of Reg

ShuffleUnit block. ShuffleNet unit with pointwise group convolution (GConv) and channel shuffle. Args: in_channels (int): The in

Generate the target heatmap. Required keys: 'joints_3d', 'joints_3d_visible', 'ann_info'. Modified keys: 'target', and 'target_weight'.

Heads for multi-stage multi-unit heads used in Multi-Stage Pose estimation Network (MSPN), and Residual Steps Networks (RSN). Args: u

ViPNAS_ResLayer to build ResNet style backbone. Args: block (nn.Module): Residual block used to build ViPNAS ResLayer. num_blocks

HourglassNet backbone. Stacked Hourglass Networks for Human Pose Estimation. More details can be found in the `paper <https://arxiv.org/a

InvertedResidual block for ShuffleNetV2 backbone. Args: in_channels (int): The input channels of the block. out_channels (int): T

Top-down pose detectors for multi-frame settings for video inputs. `"Learning temporal pose estimation from sparsely-labeled videos" <https:/

PoseWarper neck. `"Learning temporal pose estimation from sparsely-labeled videos" <https://arxiv.org/abs/1906.04016>`_. Args: i

Residual Steps Network backbone. Paper ref: Cai et al. "Learning Delicate Local Representations for Multi-Person Pose Estimation" (ECCV 2020).

ArcFace with ResNet architectures. Ref: ArcFace: Additive Angular Margin Loss for Deep Face Recognition. Args: block (str): Block us

SEBottleneck block for SEResNet. Args: in_channels (int): The input channels of the SEBottleneck block. out_channels (int): The o

Top-down pose detectors. Args: backbone (dict): Backbone modules to extract feature. keypoint_head (dict): Keypoint head to proce

Data augmentation with random image flip for bottom-up. Args: flip_prob (float): Probability of flip.

CPM backbone. Convolutional Pose Machines. More details can be found in the `paper <https://arxiv.org/abs/1602.00134>`__ . Args:

Transformer block that takes in a cross-attention condition. Designed for SparseLRM architecture.

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).

Feature fusion block.

The heatmap parser for post processing.

Hourglass-AE Network proposed by Newell et al. Associative Embedding: End-to-End Learning for Joint Detection and Grouping. More details

High-Resolution Module for LiteHRNet. It contains conditional channel weighting blocks and shuffle blocks. Args: num_branches (

Lite-HRNet backbone. `Lite-HRNet: A Lightweight High-Resolution Network <https://arxiv.org/abs/2104.06403>`_. Code adapted from 'https:/

MSPN backbone. Paper ref: Li et al. "Rethinking on Multi-Stage Networks for Human Pose Estimation" (CVPR 2020). Args: unit_channels (

Image to Patch Embedding

r"""A helper class to calculate running average in a sliding window. Args: window (int): The size of the sliding window.

Squeeze-and-Excitation Module. Args: channels (int): The input (and output) channels of the SE layer. ratio (int): Squeeze ratio

Discriminator for SMPL pose and shape parameters. It is composed of a discriminator for SMPL shape parameters, a discriminator for SMPL pose p

Swin Transformer backbone. A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` - https://

Automatically selects the most suitable model based on available GPU memory. Attributes: model_requirements (dict): Model names mappe

CpmBlock for Convolutional Pose Machine. Args: in_channels (int): Input channels of this block. channels (list): Output channels

HRFormer backbone. This backbone is the implementation of `HRFormer: High-Resolution Transformer for Dense Prediction <https://arxiv.org/abs/

HRNet backbone. `High-Resolution Representations for Labeling Pixels and Regions <https://arxiv.org/abs/1904.04514>`__ Args: ext

Loading image(s) from file. Required key: "image_file". Added key: "img". Args: to_float32 (bool): Whether to convert the loade

A torch-native trimesh class, with support for ``ply/obj/glb`` formats. Note: This class only supports one mesh with a single textur

Modulation with adaLN. References: DiT: https://github.com/facebookresearch/DiT/blob/main/models.py#L101

Code modified from https://github.com/shubham-goel/4D-Humans/blob/a0def798c7eac811a63c8220fcc22d983b39785e/hmr2/models/components/pose_transforme

Preprocessing under cv2 conventions.

ResNeXt backbone. Please refer to the `paper <https://arxiv.org/abs/1611.05431>`__ for details. Args: depth (int): Network depth