Convert axis angles to rotation 6d representations. Args: axis_angle (Union[torch.Tensor, numpy.ndarray]): input shape sh

Computes the precision@k for the specified values of k.

Add an camera to the scene. Args: camera_parameter (CameraParameter): An instance of class CameraParameter which

Applies a batch of rigid transformations to the joints. Parameters ---------- rot_mats : torch.tensor BxNx3x3 Tensor of rotation

Calculates the rotation matrices for a batch of rotation vectors Parameters ---------- rot_vecs: torch.tensor Nx3

Flip bboxes horizontally or vertically. Args: bboxes (Tensor): Shape (..., 4*k) img_shape (tuple): Image shape. direction

Convert bbox coordinates from (x1, y1, x2, y2) to (cx, cy, w, h). Args: bbox (Tensor): Shape (n, 4) for bboxes. Returns: Ten

Calculates the per vertex displacement due to the blend shapes. Parameters ---------- betas : torch.tensor Bx(num_betas) Blend sh

Useful args: - backbone: backbone name - lr_backbone: - dilation - return_interm_indices: available: [0,1,2,3], [1,2,3], [3] - ba

Build lights.

Builder of IoU calculator.

Build post processing function.

Collect results in cpu.

Compute loss for smplx left/right hand pose.

Concat images of different size.

Convert opencv 3x3 intrinsic matrix to 4x4. Args: K (Union[torch.Tensor, np.ndarray]): Input 3x3 intrinsic matrix, left mm de

Convert bbox to intrinsic parameters. Args: bbox (np.ndarray): (frame, num_person, 4), (frame, 4), or (4,) img_width (int): image

Convert BatchNorm/SyncBatchNorm in module into FrozenBatchNorm. Args: module (torch.nn.Module): Returns: If

Compute intrinsic camera matrix from orig_cam parameter of smpl. .. code-block:: python r > 1:: K = [[sx*r, 0

If a key contains 'keypoints', and f'{key}_mask' is in self.keys(), invalid zeros will be inserted to the right places and f'{key}_mask'

Convert smpl pose dict to full pose tensor. Args: smpl_dict (dict): smpl pose dict. Returns: torch: full pos

Coerces input into a random number generator. If the input is None, then a global random state is returned. If the input is a numeric value,

Flip 3d joints. Args: joints_3d (np.ndarray): joints in shape (N, 3, 2) width (int): Image width joint_pairs (list): flip

flip SMPL pose parameters.

Flip 3d xyz joints. Args: joints_3d (np.ndarray): Joints in shape (N, 3) joint_pairs (list): flip pairs for joints Returns:

Interpolate the vertex attributes to a map. Args: verts_attr (Optional[torch.Tensor]): shape should be (N, V, C),

Forward function for `Transformer`. Args: x (Tensor): Input query with shape [bs, c, h, w] where c = embed_dims.

Forward function.

Generalized IoU from https://giou.stanford.edu/ The boxes should be in [x0, y0, x1, y1] format Returns a [N, M] pairwise matrix, where N = l

Get the identity normal vector which stretchs out of the camera plane. Could pass `R` to override the camera extrinsic rotation matri

Get default hmr intrinsic, defined by how you trained. Args: num_frame (int, optional): num of frames. Defaults to 1. focal_lengt

Args: x (float): Input images(BxCxHxW) axis (int): The index for weighted mean other_axis (int): The other index Re

Get keypoints2d projected from keypoints3d. Args: device (str): Device name, should be Kinect or iPhone.

Get several frames captured by a kinect RGB camera. Args: kinect_id (int): ID of a kinect, starts from 0.

Get intrinsics of a kinect RGB camera by kinect id. Args: kinect_id (int): ID of a kinect, starts from 0.

Get resolution of a kinect RGB camera by kinect id. Args: kinect_id (int): ID of a kinect, starts from 0.

Get several frames captured by a kinect depth camera. Args: kinect_id (int): ID of a kinect, starts from 0.

Get extrinsics(cam2world) of a kinect depth camera by kinect id. Args: kinect_id (int): ID of a kinect, starts fr

Get a a matrix-type parameter by key. Args: key (str): Name of the parameter. Raises: KeyErro

Copy and rename a folder of images into a new folder following the `img_format`. Args: input_folder (str): input folder. outp

Join the meshes along the scene dim. Args: meshes (Union[ParametricMeshes, List[ParametricMeshes]]): ParametricMeshes object

Load name and parameters from a dict. Args: json_dict (dict): A dict comes from self.to_dict().

Appends a row of [0,0,0,1] to a batch size x 3 x 4 Tensor. Parameters ---------- :param input: A tensor of dimensions batch size x 3 x 4

nn.Module: the normalization layer named "norm1"

nn.Module: the normalization layer named "norm2"

This function computes the perspective projection of a set of points. Input: points (bs, N, 3): 3D points rotation (bs, 3, 3): Ca

Get item from self.human_data.

vertices: torch.Tensor of shape (N, 3), 3 stands for xyz

Obtain the mean of tensor on different GPUs.

Render SMPL or SMPL-X mesh or silhouette into differentiable tensors, and export video or images. Args: # smpl parameters: po

Rotate SMPL pose parameters by rot degrees.

This function is borrowed from https://github.com/kornia/kornia Convert 3x4 rotation matrix to Rodrigues vector Args: rotation_ma

Search the corresponding limbs following the basis human_data limbs. The mask could mask out the incorrect keypoints. Args: data_sour

Set requies_grad for all the networks. Args: nets (nn.Module | list[nn.Module]): A list of networks or a single network.

Set y reverse before add_camera if it is needed. Vedo defines y+ as up direction. When visualizing kinect cameras, y- is up, call set

This function disables printing when not in master process.

Convert the model into training mode while keep normalization layer freezed.

Resample the vertex attributes from a map. Args: maps_padded (torch.Tensor): shape should be (N, H, W, C). Required.

Args: x: (B, H, W, C) window_size (int): window size Returns: windows: (num_windows*B, window_size, window_size, C)

Check whether the key matches definition in CameraParameter.SUPPORTED_KEYS. Args: key (Any): Key in Camer

Check whether the temporal length of val matches other values. Args: key (Any): Key in HumanData. val

Check whether the shape of val matches definition in HumanData.SUPPORTED_KEYS. Args: key (Any): Key in Hu

Check whether the type of val matches definition in HumanData.SUPPORTED_KEYS. Args: key (Any): Key in Hum

Get the error message when a key fails the check. Args: key (Any): The key with wrong. Returns:

Get the warning message when a key fails the check. Args: key (Any): The key with wrong. Returns:

For each key in this HumanData, get the dimension for slicing. 0 for default, if no other value specified. Returns: dict:

Get the error message when a value fails the check. Returns: str: The error message.