Return the list of external input/output of sub-graph, each element is tuple of the name and corresponding version in predict_net. exter

Generate training targets. Args: gt_bboxes_3d (:obj:`LiDARInstance3DBoxes`): Ground truth gt boxes. gt_labels_3d (tor

Generate training targets. Args: gt_bboxes_3d (:obj:`LiDARInstance3DBoxes`): Ground truth gt boxes. gt_labels_3d (tor

Generate training targets. Args: gt_bboxes_3d (:obj:`LiDARInstance3DBoxes`): Ground truth gt boxes. gt_labels_3d (tor

Rotate nms for each task. Args: batch_cls_preds (list[torch.Tensor]): Prediction score with the shape of [N].

将tensor转化为cv2格式

https://github.com/Megvii-BaseDetection/BEVStereo/blob/master/dataset/nusc_mv_det_dataset.py#L48

Indoor Evaluation. Evaluate the result of the detection. Args: gt_annos (list[dict]): Ground truth annotations. dt_annos (li

Inference with slide/whole style. Args: points (torch.Tensor): Input points of shape [B, N, 3+C]. img_metas (list[dic

Args: predictions: return values of :meth:`forward()`. proposals (list[Instances]): proposals that match the features

Default initialization for Parameters of Module.

Dump an "Instances" object to a COCO-format json that's used for evaluation. Args: instances (Instances): img_id (int): the

Add object_descriptions and logit (if applicable) to detectron2's instances_to_coco_json

Recursively instantiate objects defined in dictionaries by "_target_" and arguments. Args: cfg: a dict-like object with "_target

Interpolate a polyline. Args: vector (array): line coordinates, shape (M, 2) num_pts (int):

Inverse function of sigmoid. Args: x (Tensor): The tensor to do the inverse. eps (float): EPS avoid numerical

Launch multi-gpu or distributed training. This function must be called on all machines involved in the training. It will spawn child proc

Load annotations from ann_file. Args: ann_file (str): Path of the annotation file. Returns: list[dict]: List

Load annotations from ann_file. Args: ann_file (str): Path of the annotation file. Returns: list[dict]: List

Load annotation from COCO style annotation file. Args: ann_file (str): Path of annotation file. Returns: lis

Load annotations from ann_file. Args: ann_file (str): Path of the annotation file. Returns: list[dict]: List

Adapt for petrel. Origin Implementation: img = Image.open(filename) copy from LoadMultiViewImageFromFiles Image.open() default is RGB

Loads object predictions from file. :param result_path: Path to the .json result file provided by the user. :param max_boxes_per_sample:

Prepare data related to Lyft dataset. Related data consists of '.pkl' files recording basic infos. Although the ground truth database and 2D

Return list of map mask layers of the specified patch. :param map_geom: List of layer names and their corresponding geometries.

Map voxel features to its corresponding points. Args: pts_coors (torch.Tensor): Voxel coordinate of each point. voxel

Map the centers of voxels to its corresponding points. Args: pts_coors (torch.Tensor): The coordinates of each points, shape

Convert pred_mask_logits to estimated foreground probability masks while also extracting only the masks for the predicted classes in pred_ins

Compute the loss between predicted and ground truth trajectories. Args: traj_output (Tensor): Predicted trajectory, shap

Compute the loss between predicted and ground truth trajectories. Args: traj_output (Tensor): Predicted trajectory, shap

r""" Args: query, key, value: map a query and a set of key-value pairs to an output. See "Attention Is All You Need" for more

r""" Args: query, key, value: map a query and a set of key-value pairs to an output. See "Attention Is All You Need" for more

r"""Apply positional encoding to the input. Args: tensor (torch.Tensor): Input tensor to be positionally encoded. encoding_size (o

NMS function GPU implementation (for BEV boxes). The overlap of two boxes for IoU calculation is defined as the exact overlapping area of the

Performs non-maximum suppression (NMS) on the rotated boxes according to their intersection-over-union (IoU). Rotated NMS iteratively re

Returns a md instance corresponding to having no predictions.

Returns a md instance corresponding to having no predictions.

Returns a md instance corresponding to having no predictions.

Returns: (H, W) array, a mask for all pixels that have a prediction

Prepare data related to nuScenes dataset. Related data consists of '.pkl' files recording basic infos, 2D annotations and groundtruth databas

Prepare data related to nuScenes dataset. Related data consists of '.pkl' files recording basic infos, 2D annotations and groundtruth databas

Obtain history BEV features iteratively. To save GPU memory, gradients are not calculated.

Obtain the info with RT matric from general sensor to Top LiDAR. Args: nusc (class): Dataset class in the nuScenes dataset. senso

Obtain the info with RT matric from general sensor to Top LiDAR. Args: nusc (class): Dataset class in the nuScenes dataset. senso

r"""Convert an integer label x-D tensor to a one-hot (x+1)-D tensor. Args: labels: tensor with labels of shape :math:`(N, *)`, where N is

Test without augmentation.

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-eye

Calculate 3D overlaps of two boxes. Note: This function calculates the overlaps between ``boxes1`` and ``boxes2``, ``

Given two lists of boxes of size N and M, compute the intersection area between __all__ N x M pairs of boxes. The box order must be (xmin

recursively (post-order) update all modules with the target type and its subclasses, make a initialization/composition/inheritance/... via th

Return true if point is in canvas

Checks whether or not the four pixels directly around the given point has less depth than the given vertex depth If True, this means that the

Obtain image features using points. Args: img_meta (dict): Meta info. img_features (torch.Tensor): 1 x C x H x W image features.

Check points is in 3d convex polygons. Args: points (np.ndarray): Input points with shape of (num_points, 3). polygon_surfaces (n

Args: polygons (list[ndarray]): each array has shape (Nx2,) height, width (int) Returns: ndarray: a bool mask of sha

Get the intersection of the convex hull of the reprojected bbox corners and the image canvas, return None if no intersection. Args: c

projects hidden states correctly to key/query states