Returns: The number of processes in the process group

Returns: The number of processes in the process group

Computes bbox detections at the given scale. Returns predictions in the scaled image space.

Compute the inverse of this transform.

Returns: True if distributed training is enabled

Returns: True if distributed training is enabled

L2-normalize columns of X

Load anns with the specified ids. If ids=None load all anns. Args: ids (int array): integer array of annotation ids Return

Load data from multiview detection dataset Args: data_dir: path to the dataset world_name: name of the world exc

Load categories with the specified ids. If ids=None load all images. Args: ids (int array): integer array of image ids Ret

Load the scene in the MoveIt! planning scene. Loads a static floor and if needed, a static wall to the MoveIt! planning scene.

Match 2D bounding boxes by projecting 3D points to 2D image planes and checking which points are inside the bounding boxes. The match between

return iou of a and b, numpy version for data augenmentation

Open the gripper.

Predicts the joint axes for the given object. Given a model name and list of types, - first get the list of joint names by calling /g

Prepare variables including APIs and objects for LMPs

Project 3D points to 2D image plane @param points_3d: 3D points in the form of (N, 3) @param intrinsic: camera intrinsic parameters @

Query extrinsic transform from camera to target frame

Check if all area of the bounding box is reachable by the panda arm.

Reduce the loss dictionary from all processes so that process with rank 0 has the averaged results. Returns a dict with the same fields as

Adds a given mesh to the scene and registers it as an object.

Reset the tsdf volume.

Reset the environment to the start state, and save the start state.

Round weighted channel number (make divisible operation). Parameters: ---------- channels : int or float Original number of

Wrapper function which calculates the results.

Select only predictions which have a `score` > self.confidence_threshold, and returns the predictions in descending order of score

Setup LMP for tabletop manipulation tasks. Args: env: the environment wrapper cfg_tabletop: the config file for tabletop man

Serialize Transform object into a dictionary.

Update the 3D bounding boxes for each object. It filters the points inside the 2D bounding boxes and then fit the 3D bounding boxes.

Update the multi-view fusion result. Args: data: Dictionary of sensor data.

Update the object instances in the scene: - object_names: list of object names - object_2d_bbox_dict: dictionary of 2

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

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

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

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

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

Convert a moveit_msgs/AllowedCollisionMatrix to a matrix of bools :param acm: moveit_msgs/AllowedCollisionMatrix :return: np.ndarray of

Arguments: num_classes (int): number of output classes input_size (int): number of channels of the input once it's fl

Arguments: proposal_matcher (Matcher) fg_bg_sampler (BalancedPositiveNegativeSampler) box_coder (BoxCoder

Arguments: in_channels (int): number of channels of the input feature num_anchors (int): number of anchors to be pred