Helper function to synchronize (barrier) among all processes when using distributed training

Update collision matrix with a list of link tuples and allowed values :param link_tuple_lists: [(link_1, link_2), ...] :par

Performs non-maximum suppression on a boxlist, with scores specified in a boxlist field via score_field. Arguments: boxlist(BoxL

Download images from mscoco.org server. Args: save_dir (str): dir to save downloaded images img_ids (int array): img i

Run per image evaluation on given images and store results (a list of dict) in self.eval_imgs.

Get object bounding box.

Get ground truth object pose from gazebo

Caffe2 implementation uses XavierFill, which in fact corresponds to kaiming_uniform_ in PyTorch

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

Move the robot to a given pose with given orientation.

Compute and display summary metrics for evaluation results.

Arguments: pre_nms_top_n (int) post_nms_top_n (int) nms_thresh (float) min_size (int)

Arguments: pre_nms_top_n (int) post_nms_top_n (int) nms_thresh (float) min_size (int)

Update default collision entries with a list of links and allowed values :param link_lists: [link_1, link_2, ...] :param all

Calculate TF-same like padding size. Parameters: ---------- x : tensor Input tensor. kernel_size : int Convoluti

Efficient version of torch.cat that avoids a copy if there is only a single element in a list

1x1 version of the standard convolution block. Parameters: ---------- in_channels : int Number of input channels. out_ch

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

Get ann ids that satisfy given filter conditions. Args: img_ids (int array): get anns for given imgs cat_ids (int arra

Chooses static padding if you have specified an image size, and dynamic padding otherwise. Static padding is necessary for ONNX exporting of

get sensor data from perception model

Executes a grasp at a given pose with given orientation. Args: position: The position of the grasp. orient

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

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

Load the selected models from the metadata file.

Filter open3d point cloud with frustum filters by projecting 3D points onto 2D image planes and checking if they are within the 2D bounding

Open the gripper.

very similar to the smooth_l1_loss from pytorch, but with the extra beta parameter

/////////////////////////////////////////////////////////////////////////// Load the controller

Arguments: in_channels_list (list[int]): number of channels for each feature map that will be fed out

Arguments: in_channels_list (list[int]): number of channels for each feature map that will be fed out

Arguments: cfg : config in_channels (int): number of channels of the input feature num_ancho

Construct a layernorm module in the TF style (epsilon inside the square root).

Class for reading and visualizing annotations. Args: annotation_path (str): location of annotation file

Efficient version of torch.cat that avoids a copy if there is only a single element in a list

Validates model name.

Create query string for the remote service. Args: question: String, question text image_list: List, list of i