Initializes DDNDeepLabV3 model Args: backbone_name: string, ResNet Backbone Name [ResNet50/ResNet101]

Initializes 2D convolution collapse module Args: model_cfg: EasyDict, Model configuration grid_size: (X, Y, Z

:param radius: float, radius of ball :param nsample: int, maximum number of features to gather in the ball :param use_xyz:

:param npoint: int :param radii: list of float, list of radii to group with :param nsamples: list of int, number of samples i

:param mlp: list of int :param bn: whether to use batchnorm

Return an activation function given a string

Cosine anneal from `start` to `end` as pct goes from 0.0 to 1.0.

Args: gt_boxes: (B, M, 8) Returns:

Args: all_anchors: [(N, 7), ...] gt_boxes: (B, M, 8) Returns:

Args: anchors: [(N, 7), ...] gt_boxes: (B, M, 8) Returns:

:param ctx: :param grad_out: (B, C, N) tensor with gradients of outputs :return: grad_features: (B, C, M) tensor

:param ctx: :param grad_out: (B, C, npoint, nsample) tensor of the gradients of the output from forward :return:

Args: ctx: grad_out: (M1 + M2 ..., C, nsample) tensor of the gradients of the output from forward Returns:

Args: ctx: grad_out: (N1 + N2 ..., C) Returns: grad_features: (M1 + M2 ..., C)

Args: ctx: grad_new_features: (M1 + M2 ..., num_c_out), num_c_out = num_c_out_each_grid * num_total_grids Re

:param grad_out: (N, out_x, out_y, out_z, C) :return: grad_in: (npoints, C)

:param ctx: :param grad_out: (B, C, N) tensor with gradients of outputs :return: grad_features: (B, C, M) tensor

:param ctx: :param grad_out: (B, C, npoint, nsample) tensor of the gradients of the output from forward :return:

r""" Parameters ---------- grad_out : torch.Tensor (B, c, n) tensor with gradients of ouputs Returns

r""" Parameters ---------- grad_out : torch.Tensor (B, C, npoint, nsample) tensor of the gradients of the output

Converts depth map into bin indices Args: depth_map: (H, W), Depth Map mode: string, Discretiziation mode (See https://arxiv.

Returns the four bottom corners. :return: <np.float: 3, 4>. Bottom corners. First two face forward, last two face backwards.

:param boxes3d: (N, 7) [x, y, z, l, h, w, r] in rect camera coords :param calib: :return: box_2d_preds: (N, 4) [x1, y1, x2, y2]

Args: boxes3d_camera: (N, 7) [x, y, z, l, h, w, r] in rect camera coords calib: Returns: boxes3d_lidar: [x, y, z, dx

Args: boxes3d_lidar: (N, 7) [x, y, z, dx, dy, dz, heading], (x, y, z) is the box center Returns: boxes3d_fakelidar: [x, y, z

:param boxes3d_lidar: (N, 7) [x, y, z, dx, dy, dz, heading], (x, y, z) is the box center :param calib: :return: boxes3d_camera: (

Args: boxes_a: (N, 7) [x, y, z, dx, dy, dz, heading] boxes_b: (N, 7) [x, y, z, dx, dy, dz, heading] Returns: