Args: root_path: dataset_cfg: class_names: training: logger:

: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

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

Args: query_ranges: list of int, list of neighbor ranges to group with nsamples: list of int, number of samples in ea

Args: radius: float, radius of ball nsample: int, maximum number of features to gather in the ball use_xy

Args: mlp: list of int

Args: mlp: num_voxels: max_neighbour_distance: neighbor_type: 1: ball, others: cube

Args: gamma: Weighting parameter to balance loss for hard and easy examples. alpha: Weighting parameter to balance lo

Args: beta: Scalar float. L1 to L2 change point. For beta values < 1e-5, L1 loss is computed.

Args: root_path: dataset_cfg: class_names: training: logger:

Initializes convolutional block Args: in_channels: int, Number of input channels out_channels: int, Number of

Initializes Grid Generator for frustum features Args: grid_size: [X, Y, Z], Voxel grid size pc_range: [x_min,

Initializes module Args: mode: string, Sampling mode [bilinear/nearest] padding_mode: string, Padding mode fo

Initializes module to transform frustum features to voxel features via 3D transformation and sampling Args: model_cfg: Ea

Initialize frustum feature network via depth distribution estimation Args: model_cfg: EasyDict, Depth classification netw

Initialize fixed foreground/background loss balancer Args: fg_weight: float, Foreground loss weight bg_weight

Initializes DDNLoss module Args: weight: float, Loss function weight alpha: float, Alpha value for Focal Loss

Initializes depth distribution network. Args: constructor: function, Model constructor feat_extract_layer: st