Private function to load point clouds data. Args: pts_filename (str): Filename of point clouds data. Returns:

Private function to load point clouds data. Args: pts_filename (str): Filename of point clouds data. Returns:

Private function to load point clouds data. Args: pts_filename (str): Filename of point clouds data. Returns:

Private function to load point clouds data. Args: pts_filename (str): Filename of point clouds data. Returns:

Private function to load point clouds data. Args: pts_filename (str): Filename of point clouds data. Returns:

Private function to load 3D semantic segmentation annotations. Args: results (dict): Result dict from :obj:`mmdet3d.CustomData

Per scene loss function. Args: center_preds (list[list[Tensor]]): Centerness predictions for all scenes.

Calculate regularization loss for model weight in training.

Per scene loss function. Args: center_preds (list[Tensor]): Centerness predictions for all levels. bbox_preds (li

Per scene loss function. Args: center_preds (list[Tensor]): Centerness predictions for all levels. bbox_preds (li

Make a convolutional block. Args: in_channels (int): Number of input channels. out_channels (int): Number of outp

Construct Conv-Norm-Act block. Args: in_channels (int): Number of input channels. out_channels (int): Number of o

Conv modules. Args: in_channels (int): Input feature channel. out_channels (int): Output feature channel.

Make a layer from several residual blocks. Args: stride (int): Stride of the first residual block. n_channels (in

Make upsampling convolutional block. Args: in_channels (int): Number of input channels. out_channels (int): Numbe

Construct DeConv-Norm-Act-Conv-Norm-Act block. Args: in_channels (int): Number of input channels. out_channels (i

Pad images according to ``self.size``.

Patch points sampling. First sample a valid patch. Then sample points within that patch to a certain number. Args:

Args: points (np.ndarray): Input points with shape of (num_points, 3). polygon_surfaces (np.ndarray): Polygon surfaces with sh

Points random sampling. Sample points to a certain number. Args: points (np.ndarray | :obj:`BasePoints`): 3D Points

convert kitti points(N, >=3) to voxels. Args: points (np.ndarray): [N, ndim]. points[:, :3] contain xyz points and points

convert kitti points(N, >=3) to voxels. Args: points (np.ndarray): [N, ndim]. points[:, :3] contain xyz points and points

Perform feature aggregation using pooling operation. Args: features (torch.Tensor): (B, C, N, K) Features of

Prunes the tensor by score thresholding. Args: x (SparseTensor): Tensor to be pruned. scores (SparseTensor): Scor

Randomly get another item with the same flag. Returns: int: Another index of item with the same flag.

Randomly get another item with the same flag. Returns: int: Another index of item with the same flag.

Private function to randomly set the scale factor. Args: input_dict (dict): Result dict from loading pipeline. Retu

Reset the index of batchsampler to zero.

Private function to rotate bounding boxes and points. Args: input_dict (dict): Result dict from loading pipeline. R

Private function to rotate bounding boxes and points. Args: input_dict (dict): Result dict from loading pipeline.

Get the 3D rotation matrix. Args: rot_mat_T (np.ndarray): Transposed rotation matrix. angle (float): Rotation angle.

Sample specific number of ground truths and return indices. Args: num (int): Sampled number. Returns:

Randomly sample some negative samples.

Randomly sample some positive samples.

Private function to scale bounding boxes and points. Args: input_dict (dict): Result dict from loading pipeline. Re

Train semantic head. Args: x (torch.Tensor): Point-wise semantic features for segmentation voxels_dict (dict): Co

Set flag according to image aspect ratio. Images with aspect ratio greater than 1 will be set as group 1, otherwise group 0. In 3D

Multi-class nms for a single scene. Args: bboxes (Tensor): Predicted boxes of shape (N_boxes, 6) or (N_boxes,

Multi-class nms for a single scene. Args: bboxes (Tensor): Predicted boxes of shape (N_boxes, 6) or (N_boxes,

Sampling points in a sliding window fashion. First sample patches to cover all the input points. Then sample points in each patch

Get indexes based on scores. Args: scores (torch.Tensor): scores with the shape of [B, N, W, H]. K (int, optional

Private function to translate bounding boxes and points. Args: input_dict (dict): Result dict from loading pipeline.

Private function to translate points. Args: input_dict (dict): Result dict from loading pipeline. trans_factor (n

Add extra files that are required to support MIM into the package. These files will be added by creating a symlink to the originals if the

Convert the rotation difference to difference in sine function. Args: boxes1 (torch.Tensor): Original Boxes in shape (NxC), wh

Maps predictions to ScanNet evaluator format. Args: masks (list[torch.Tensor]): Per scene predicted instance masks. labels (l

Assign gt and predicted instances for a single scene. Args: pred_info (dict): Predicted masks, labels and scores. gt_ids (np.

Pre-compute features with weight matrices in weight bank. This function is used before cuda op assign_score_withk in CUDA version PAConv.

Perform weighted sum to aggregate output features according to scores. This function is used in non-CUDA version of PAConv. Compared to th

Test with augmentations. Args: points (list[torch.Tensor]): List of points of shape [B, N, 3+C]. img_metas (list[

Test function with augmentaiton.

r"""Test function with augmentation, image network pretrain. May refer to `<https://github.com/open-mmlab/mmdetection/blob/master/mmdet/model

Test function of point cloud branch with augmentaiton. The function implementation process is as follows: - step 1: map feat

Test function of point cloud branch with augmentaiton.

Test function of point cloud branch with augmentaiton.

Calculate auxiliary loss. Args: points (torch.Tensor): Mean feature value of the points. point_cls (torch.Tensor)

Calculate average precision (for single or multiple scales). Args: recalls (np.ndarray): Recalls with shape of (num_scales, num_dets)

Calculate overlap between two set of axis aligned 3D bboxes. If ``is_aligned`` is ``False``, then calculate the overlaps between each bbox o

Calculate the IoU loss (1-IoU) of two sets of axis aligned bounding boxes. Note that predictions and targets are one-to-one corresponded.

Map bboxes from testing scale to original image scale. Args: bboxes (:obj:`BaseInstance3DBoxes`): Boxes to be mapped back. sc

Transform 2d bbox according to img_meta. Args: img_meta(dict): Meta info regarding data transformation. bbox_2d (torch.Tensor

Calculate 3D IoU using cuda implementation. Note: This function calculates the IoU of 3D boxes based on their volumes. IoU ca

Transform 3D boxes. Args: boxes (np.ndarray): 3D boxes to be transformed. loc_transform (np.ndarray): Location transform to b

Build data config for loading visualization data.

Build roi layers using `layer_cfg`

Build roi layers using `layer_cfg`

Build segmentor.

Calculate the Euclidean distance between two sets of points. Args: xyz1 (torch.Tensor): (N, 3), the first set of points. xyz2

Fast iou algorithm. this function can be used independently to do result analysis. Must be used in CAMERA coordinate system. Args:

Find all boxes in which each point is, as well as the offsets from the box centers. Args: points (torch.Tensor): [M,

Convert transformation matrix in Cartesian coordinates to homogeneous format. Args: mat (np.ndarray): Transformation

Center to minmax. Args: centers (np.ndarray): Center points. dims (np.ndarray): Dimensions. origin (list or array or

Whether the mean of points to plane distance is lower than thresh. Args: plane_equ (torch.Tensor): Plane to be checked.

Check whether is a horizontal plane. Args: points (torch.Tensor): Points of input. Returns: Bool: Flag

Circular NMS. An object is only counted as positive if no other center with a higher confidence exists within a radius r using a bird

Class agnostic nms for single batch. Args: mlvl_bboxes (torch.Tensor): Bboxes from Multi-level. mlvl_bboxes_for_n

Class agnostic nms. Args: obj_scores (torch.Tensor): Objectness score of bounding boxes. sem_scores (torch.Tensor

Sigmoid function for input feature. Args: x (torch.Tensor): Input feature map with the shape of [B, N, H, W]. eps (float, opt

Combine predictions in waymo format for each sample together. Args: pathnames (str): Paths to save predictions. Ret

Combine all the prediced depths with depth uncertainty. Args: depth (torch.Tensor): Predicted depths of each object.

Deprecated sample_per_gpu in cfg.data.

Averages AP scores for all categories. Args: aps (np.array): AP scores for all thresholds and categories. options (dict): Sca

Compute loss of primitive module. Args: primitive_center (torch.Tensor): Pridictions of primitive center. primiti

Concat data_infos from several datasets to form self.data_infos. Args: data_infos (list[list[dict]])

Concat scene_idxs from several datasets to form self.scene_idxs. Needs to manually add offset to scene_idxs[1, 2, ...]. Args:

Convert keys in pycls pretrained RegNet models to mmdet style.

Convert boxes or points from `src` mode to `dst` mode. Args: input (tuple | list | np.ndarray | torch.Tensor |

Convert action.

Convert config's naiveSyncBN to BN. Args: config (str or :obj:`mmcv.Config`): Config file path or the config object.