Register the pybullet body with the tracker so we can obtain updated simulator states, map them to the correct mesh asset, and send t

Convert rotations given as rotation matrices to quaternions. Args: matrix: Rotation matrices as tensor of shape (..., 3, 3). Re

Returns a set of camera config parameters Returns: YACS CfgNode: Cam config params

input: B N 3 --------------------------- output: B G M 3 center : B G 3

Return the end effector pose. Returns: 4-element tuple containing - np.ndarray: x, y, z position of the EE

Move the end effector to the specifed pose. Args: pos (list or np.ndarray): Desired x, y, z positions in the robot's

Creates the output 2-layer MLPs used for predicting translation and rotation Args: pooled_dim (int): Input dimen

Uses the task name to call the respective method for generating a pose that's likely to satisfy the relation

Normalize coordinate to [0, 1] for unit cube experiments. Corresponds to our 3D model Args: x (tensor): coordinate reso

Move the arm to the specified joint position(s). Args: position (float or list): desired joint position(s).

Compute the inverse kinematics solution given the position and orientation of the end effector. Args: pos (list

Convert rotations given as rotation matrices to axis/angle. Args: matrix: Rotation matrices as tensor of shape (..., 3, 3). Ret

Normalize coordinate to [0, 1] for unit cube experiments Args: p (tensor): point padding (float): conventional padding paramter

Decay the learning rate with half-cycle cosine after warmup

Convert 3d vector of axis-angle rotation to 4x4 rotation matrix Args: angle_axis (Tensor): tensor of 3d vector of axis-angle rotations.

input: B N 3 --------------------------- output: B G M 3 center : B G 3

Furthest point sampling. data B N 3 number int batch B*N 1 (batch index for each point in the flattened input)

Return the joint position(s) of the arm. Args: joint_name (str, optional): If it's None, it will return

:param self: :param x: Shape (batch_size, num_points) :param y: :param z: :param intrinsics: :return:

Pre-processes the input point cloud and per-point features. Consists of scaling the 3D point clouds to be roughly ~unit bounding box

Multiply two quaternions. Usual torch rules for broadcasting apply. Args: a: Quaternions as tensor of shape (..., 4), real part

Arguments: in_channels: int, number of channels in the input tensor. Default is 3 for RGB images. dep

Args: feat_dim (int): Dimensionality of the input features (typically 3D for point clouds, or 5D for point clouds A and B - with

Extract the first or third Euler angle from the two members of the matrix which are positive constant times its sine and cosine. Args:

Return an activation function given a string

Internal function to create the full name of the skill that was used to transition to this node (combining the skill type and the sur

Convert rotations given as axis/angle to rotation matrices. Args: axis_angle: Rotations given as a vector in axis angle form,

Args: index (int): Dataset index to sample embed_invalid (bool): If True, embed into interactive termainl if we hit a

Decay arbitrary value with half-cycle cosine after warmup

Args: bl (float): Base length bw (float): Base width bt (float): Base thickness n_shelves (int): Number of shelve

Args: bl (float): Base length bw (float): Base width bt (float): Base thickness wt (float): Wall thickness

Convert rotations given as rotation matrices to Euler angles in radians. Args: matrix: Rotation matrices as tensor of shape (..., 3,

Preprocess point cloud by applying a voxel-based downsampling, estimating the point cloud normals, and computing the FPFH features that are us

Convert rotations given as quaternions to rotation matrices. Args: quaternions: quaternions with real part first, as ten

Reset the state of the joint. It's best only to do this at the start, while not running the simulation. It will overrides all

Move the arm with the specified joint velocity(ies). Args: velocity (float or list): desired joint velocity(ies).