Decorator to make all processes in distributed training wait for each local_master to do something.

Args: radii: list of float, list of radii to group with nsamples: list of int, number of samples in each ball query

This function finds the road segment that a given location is part of, returning the edge it belongs to

Args: gt_boxes: (B, M, 8) range_image_polar: (B, 3, H, W) feature_map_size: (2) [H, W] spatia

Convert a carla location to a ROS vector3 Considers the conversion from left-handed system (unreal) to right-handed system :param c

Set config keys via list (e.g., from command line).

Check that vector (b - a) is orthogonal to the vector (c - a).

Create an `optim.Optimizer` from `opt_func` with `lr`. Set lr on `layer_groups`.

Args: box_encodings: (B, N, 7 + C) or (N, 7 + C) [x, y, z, dx, dy, dz, heading or *[cos, sin], ...] anchors: (B, N, 7

Args: boxes: (N, 7 + C) [x, y, z, dx, dy, dz, heading, ...] anchors: (N, 7 + C) [x, y, z, dx, dy, dz, heading or *[co

Args: batch_size: rois: (B, N, 7) cls_preds: (BN, num_class) box_preds: (BN, code_size)

:param pts_rect: (N, 3) :return pts_img: (N, 2)

:param mlp: list of int, spec of the pointnet before the global max_pool :param npoint: int, number of features :param radius

Args: points: (N1 + N2 + N3 + ..., 4) [bs_idx, x, y, z] gt_boxes: (B, M, 8) extend_gt_boxes: [B, M, 8]

Return intersection-over-union (Jaccard index) of boxes. Both sets of boxes are expected to be in (x1, y1, x2, y2) format. Arguments:

Args: boxes3d_fakelidar: (N, 7) [x, y, z, w, l, h, r] in old LiDAR coordinates, z is bottom center Returns:

Convert a carla rotation to rotation matrix. Considers the conversion from left-handed system (unreal) to right-handed system. Consi

:param pts: (N, 3 or 2) :return pts_hom: (N, 4 or 3)

:param pts: (N, 3 or 2) :return pts_hom: (N, 4 or 3)

:param corners3d: (N, 8, 3) :param fig: :param color: :param line_width: :param cls: :param tag: :param max_num: :ret

Overwrites the throttle a brake values of a control to perform an emergency stop. The steering is kept the same to avoid going out of

Kitti eval. support 2d/bev/3d/aos eval. support 0.5:0.05:0.95 coco AP. Args: gt_annos: dict, must from get_label_annos() in kitti_common.p

Kitti eval. support 2d/bev/3d/aos eval. support 0.5:0.05:0.95 coco AP. Args: gt_annos: dict, must from get_label_annos() in kitti_common.p

Kitti eval. support 2d/bev/3d/aos eval. support 0.5:0.05:0.95 coco AP. Args: gt_annos: dict, must from get_label_annos() in kitti_common.p

Args: data_dict: points: (N, 3 + C_in) gt_boxes: optional, (N, 7) [x, y, z, dx, dy, dz, heading]

Get actor velocity in Vector3D(vx, vy, vz). It it in Right-Hand coordinate. :return: Vector3D(vx, vy, vz)

Create a new `OptimWrapper` from `self` with another `layer_groups` but the same hyper-parameters.

Runs Non-Maximum Suppression (NMS) on inference results Returns: list of detections, on (n,6) tensor per image [xyxy, conf, cls]