Parse bbox annotation. Args: img_info (list[dict]): Image info. ann_info (list[dict]): Annotation info of an image.

Transform inputs for decoder. Args: inputs (list[Tensor]): List of multi-level img features. Returns:

Computes one-to-one matching based on the weighted costs. This method assign each query prediction to a ground truth or background. Th

Test function with augmentaiton.

GPU version of backward function. Args: grad_output (Tensor): Gradient of output tensor of forward.

GPU version of backward function. Args: grad_output (Tensor): Gradient of output tensor of forward.

Evaluate the model only at the start of training by epoch.

Binary Cross entropy loss logits: [B, H, W] Variable, logits at each pixel (between -\infty and +\infty) labels: [B, H, W] Tensor, bi

Classify each pixel.

Because I move the compute count matrix process to the `forward_test` function in the model, this function will not be used.

Convert action for single file. Args: file_idx (int): Index of the file to be converted.

Convert action for single file. Args: file_idx (int): Index of the file to be converted.

convert kitti info v1 to v2 if possible. Args: info (dict): Info of the input kitti data. - image (dict): image info

Create indoor information file. Get information of the raw data and save it to the pkl file. Args: data_path (str): Path of the data

Create info file of KITTI dataset. Given the raw data, generate its related info file in pkl format. Args: data_path (str): Path of

Create info file of lyft dataset. Given the raw data, generate its related info file in pkl format. Args: root_path (str): Path of t

Create info file of nuscene dataset. Given the raw data, generate its related info file in pkl format. Args: root_path (str): Path o

Create info file of nuscene dataset. Given the raw data, generate its related info file in pkl format. Args: root_path (str): Path o

Create info file of nuscene dataset. Given the raw data, generate its related info file in pkl format. Args: root_path (str): Path o

Create reduced point clouds for training/validation/testing. Args: data_path (str): Path of original data. pkl_prefix (str): Pref

Create info file of waymo dataset. Given the raw data, generate its related info file in pkl format. Args: data_path (str): Path of

Encode bitmap mask to RLE code. Semantic Masks only Args: mask_results (list | tuple[list]): bitmap mask results. In mask s

Test model with multiple gpus. This method tests model with multiple gpus and collects the results under two different modes: gpu and cpu mo

Evaluation. Args: voxel_semantics (torch.Tensor): semantic occpuancy ground truth. voxel_semantics_preds

Evaluation. Args: voxel_semantics (torch.Tensor): semantic occpuancy ground truth. voxel_semantics_preds (tor

Args: voxel_semantic: bs, w, h, z other four are the same

This function will be called by `tools/test.py` to evaluate the results. Args: occ_results (List[Dict]): the results of t

Evaluation in KITTI protocol. Args: results (list[dict]): Testing results of the dataset. metric (str | list[str]): M

Evaluation in nuScenes protocol. Args: results (list[dict]): Testing results of the dataset. metric (str | list[str

Performs the actual evaluation. :return: A tuple of high-level and the raw metric data.

Check if a box is visible in images but not all corners in image . :param box: The box to be checked. :param intrinsic: <float: 3, 3>. In

Export 2d annotation from the info file and raw data. Args: root_path (str): Root path of the raw data. info_path (str): Path of

Export 2d annotation from the info file and raw data. Args: root_path (str): Root path of the raw data. info_path (str): Path of

Export 2d annotation from the info file and raw data. Args: root_path (str): Root path of the raw data. info_path (str): Path of

Export 2d annotation from the info file and raw data. Args: root_path (str): Root path of the raw data. info_path (str): Path of

Export 2d annotation from the info file and raw data. Args: root_path (str): Root path of the raw data. info_path (str): Path of

Extra repr.

Extract features of points.

Applies filtering to boxes. Distance, bike-racks and points per box. :param nusc: An instance of the NuScenes class. :param eval_boxes: A

Forward function for `Detr3DTransformer`. Args: mlvl_feats (list(Tensor)): Input queries from different level.

Forward function for `Detr3DTransformerDecoder`. Args: query (Tensor): Input query with shape `(num_query, bs,

Forward Function of MultiScaleDeformAttention. Args: query (Tensor): Query of Transformer with shape (num_que

Forward function.

Forward function.

Forward function.

Forward function.

Forward function.

Forward function.

Forward function for `TransformerDecoder`. Args: bev_query (Tensor): Input BEV query with shape `(num_query, bs

Forward function for `TransformerDecoderLayer`. **kwargs contains some specific arguments of attentions. Args: quer

Forward Function of MultiScaleDeformAttention. Args: query (Tensor): BEV query of Transformer with shape (bs, num_query, embed_d

Args: bev_list (list[torch.Tensor]): each has shape (bs, h*w, embed_dims). img_metas_list (list[dict]): Include curr

Forward Function of MultiScaleDeformAttention. Args: query (Tensor): Query of Transformer with shape (num_que

Forward function for `Detr3DTransformer`. Args: mlvl_feats (list(Tensor)): Input queries from different level.

GPU version of multi-scale deformable attention. Args: value (Tensor): The value has shape (bs, num_keys, mum

GPU version of multi-scale deformable attention. Args: value (Tensor): The value has shape (bs, num_keys, mum

Forward function for `Detr3DTransformer`. Args: mlvl_feats (list(Tensor)): Input queries from different level.

Forward function for `Detr3DTransformer`. Args: multi_level_feats (list(torch.Tensor)): Current multi level img features

Forward function for `TransformerDecoder`. Args: bev_query (Tensor): Input BEV query with shape `(num_query, bs

Forward function for `TransformerDecoderLayer`. **kwargs contains some specific arguments of attentions. Args: quer

Forward function for `Detr3DTransformer`. Args: multi_level_feats (list(torch.Tensor)): Current multi level img featu

Forward Function of Detr3DCrossAtten. Args: query (Tensor): Query of Transformer with shape (num_query, bs, emb

Forward Function of MultiScaleDeformAttention. Args: query (Tensor): Query of Transformer with shape ( bs, num_

Forward function for `TransformerDecoder`. Args: bev_query (Tensor): Input BEV query with shape `(num_query, bs

Forward function for `TransformerDecoderLayer`. **kwargs contains some specific arguments of attentions. Args: quer

Forward function for `Detr3DTransformer`. Args: mlvl_feats (list(Tensor)): Input queries from different level.

Forward function for `TransformerDecoderLayer`. **kwargs contains some specific arguments of attentions. Args: query (T

Encoder of bevformer, which contains multiple layers. It can construct BEV features from flattened multi level image features. Args

Forward function for `TransformerDecoderLayer`. Args: query (Tensor): The input BEV query with shape (bs, num_queries,

Forward function for `TransformerDecoder`. Args: bev_query (Tensor): Input BEV query with shape `(num_query, bs

Forward pass of the `CostVolume` temporal fusion method. Args: bev_list (list[torch.Tensor]): each has shape (bs, h*w,

Encoder of bevformer, which contains multiple layers. It can construct BEV features from flattened multi level image features. Args

Forward function for `TransformerDecoderLayer`. Args: query (Tensor): The input BEV query with shape (bs, num_queries,

Forward Function of Detr3DCrossAtten. Args: query (Tensor): BEV query of Transformer with shape (bs, num_query, embed_dims).

Forward Function of MultiScaleDeformAttention. Args: query (Tensor): Query of Transformer with shape ( bs, num_