img: [H, W, 3] mask: [H, W]

Check if norm layer is in correct train state.

dot product (along the last dim). Args: x (Union[Tensor, ndarray]): x, [..., C] y (Union[Tensor, ndarray]): y, [..., C] Retu

Encode images with backbone and decode into a depth estimation map of the same size as input.

Encode images with backbone and decode into a depth estimation map of the same size as input.

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

Extract the target frame from 2D pose results, and pad the sequence to a fixed length. Args: pose_results (list[list[dict]]): Multi-f

r"""Convolve a tensor with a 2d kernel. The function applies a given kernel to a tensor. The kernel is applied independently at each depth ch

Converts all `BatchNorm2d` and `SyncBatchNorm` layers of provided module into `FrozenBatchNorm2d`. If `module` is itself an instance of eithe

Generate a SMPL model weight file to initialize SMPL model, and generate a 3D joints regressor file.

Get detection results of the message. Args: tag (str, optional): If given, only the results with the tag will be

Compute the focal length of the camera lens by assuming a certain FOV for the entire image Args: - fov: float, expressed in degree

Transform the grouped joints back to the image. Args: grouped_joints (list): Grouped person joints. center (np.ndarray[2, ]): Cen

transform_mat_neutral_pose is function to warp pre-defined posed to zero-pose

Draw 3D keypoints and links in 3D coordinates. Args: pose_result (list[dict]): 3D pose results containing: - "keypoints_3d" (

Inference with slide/whole style. Args: img (Tensor): The input image of shape (N, 3, H, W). img_meta (dict): Image i

Inference with slide/whole style. Args: img (Tensor): The input image of shape (N, 3, H, W). img_meta (dict): Image i

Inference function. Returns: output_regression (np.ndarray): Output regression. Args: x (torch.Tensor[N, K,

initialize network's weights init_type: normal | xavier | kaiming | orthogonal https://github.com/junyanz/pytorch-CycleGAN-an

Weight initialization for model.

Initialize model weights.

Initialize the weights.

lock modules Args: unlocked_groups (int): leave last n layer groups unlocked (default: 0)

r"""Normalize both derivative and smoothing kernel.

OKS NMS implementations. Args: kpts_db: keypoints. thr: Retain overlap < thr. sigmas: standard deviation of keypoint labe

Force to put an item into the buffer. If the buffer is already full, the earliest item in the buffer will be remove to make room for

Report timing information. Returns: dict: The key is the timer name and the value is the \ corresponding average

Set the frame image to the message.

Create configs and perform basic setups.

r"""Compute the first order image derivative in both x and y using a Sobel operator. .. image:: _static/img/spatial_gradient.png Args:

Split multi-stage outputs into heatmaps & tags. Args: outputs (list(Tensor)): Outputs of network num_joints (int): Number of join

Timing a code snippet with an assigned name. Args: timer_name (str): The unique name of the interested code snippet to