Rethinking Atrous Convolution for Semantic Image Segmentation. This head is the implementation of `DeepLabV3 <https://arxiv.org/abs/1706.05587

Atrous Spatial Pyramid Pooling (ASPP) Module. Args: dilations (tuple[int]): Dilation rate of each layer. in_channels (int): Input

Dummy class for backward-compatibility. Same as DetectionEval.

ScanNet data. Generate scannet infos for scannet_converter. Args: root_path (str): Root path of the raw data. split (str): S

KITTI predictions to Waymo converter. This class serves as the converter to change predictions from KITTI to Waymo format. Args: k

SUNRGBD data. Generate scannet infos for sunrgbd_converter. Args: root_path (str): Root path of the raw data. split (str): S

ScanNet dataset used to generate infos for semantic segmentation task. Args: data_root (str): Root path of the raw data. ann_file

Basic convolutional block for UNet. This module consists of several plain convolutional layers. Args: in_channels (int): Number of i

Pardon this horrendous name. Basically, we want every sample to be from its own group. If batch size is 4 and # of GPUs is 8, each sample of

S3DIS data. Generate s3dis infos for s3dis_converter. Args: root_path (str): Root path of the raw data. split (str): Set spl

S3DIS dataset used to generate infos for semantic segmentation task. Args: data_root (str): Root path of the raw data. ann_file (

Upsample convolution block in decoder for UNet. This upsample convolution block consists of one upsample module followed by one convolution bl

r"""Implements AdamW algorithm. Solve the bug of torch 1.8 The original Adam algorithm was proposed in `Adam: A Method for Stochastic Optimizatio

BBox3DL1Cost. Args: weight (int | float, optional): loss_weight

Attention with both self and cross Implements the decoder in DETR transformer. Args: return_intermediate (bool): Whether to r

Attention with both self and cross Implements the decoder in DETR transformer. Args: return_intermediate (bool): Whether to r

Attention with both self and cross Implements the decoder in DETR transformer. Args: return_intermediate (bool): Whether to r

Attention with both self and cross Implements the decoder in DETR transformer. Args: return_intermediate (bool): Whether to r

Attention with both self and cross Implements the decoder in DETR transformer. Args: return_intermediate (bool): Whether to r

Implements decoder layer in DETR transformer. Args: attn_cfgs (list[`mmcv.ConfigDict`] | list[dict] | dict )): Configs for

Implements decoder layer in DETR transformer. Args: attn_cfgs (list[`mmcv.ConfigDict`] | list[dict] | dict )): Configs for

Base class for BaseDecodeHead. Args: in_channels (int|Sequence[int]): Input channels. channels (int): Channels after modules,

Implements the Detr3D transformer. Args: as_two_stage (bool): Generate query from encoder features. Default: False.

Implements the Detr3D transformer. Args: as_two_stage (bool): Generate query from encoder features. Default: False. nu

Base class of Multi-modality VoxelNet.

Cross Entropy Loss with additional OHEM. Args: top_ratio (float): top ratio to be mined. Default: 0.3. top_weight (float): s

Collect data from the loader relevant to the specific task. This is usually the last stage of the data loader pipeline. Args: ke

Default formatting bundle. It simplifies the pipeline of formatting common fields for voxels, including "proposals", "gt_bboxes", "gt_labels

r"""Feature Pyramid Network. This is an implementation of paper `Feature Pyramid Networks for Object Detection <https://arxiv.org/abs/1612.03

An attention module used in Deformable-Detr. `Deformable DETR: Deformable Transformers for End-to-End Object Detection. <https://arxiv.org

r"""NuScenes Dataset. This datset only add camera intrinsics and extrinsics to the results.

r"""Monocular 3D detection on NuScenes Dataset. This class serves as the API for experiments on the NuScenes Dataset. Please refer to `NuSce

Waymo Dataset. This class serves as the API for experiments on the Waymo Dataset. Please refer to `<https://waymo.com/open/download/>`_for d

Implements the decoder in DETR3D transformer. Args: return_intermediate (bool): Whether to return intermediate outputs. coder_n

Sampler that restricts data loading to a subset of the dataset. It is especially useful in conjunction with :class:`torch.nn.parallel.Distribu

# basic logic input_sequence = [a, b, c] # given a sequence of samples prev_bev = None for each in input_sequcene[:-

Cross Entropy Loss with additional OHEM. Args: use_sigmoid (bool): Only support sigmoid in FocalLoss. Must be True. gamma (f

Computes one-to-one matching between predictions and ground truth. This class computes an assignment between the targets and the predictions b

Implements the Detr3D transformer. Args: as_two_stage (bool): Generate query from encoder features. Default: False.

Implements the Detr3D transformer. Args: as_two_stage (bool): Generate query from encoder features. Default: False.

r"""Lift-Splat-Shoot view transformer with BEVPoolv2 implementation. Please refer to the `paper <https://arxiv.org/abs/2008.05711>`_ and

Position embedding with learnable embedding weights. Args: num_feats (int): The feature dimension for each position along

Position embedding with learnable embedding weights. Args: num_feats (int): The feature dimension for each position along

Load multi channel images from a list of separate channel files. Expects results['img_filename'] to be a list of filenames. note that

Load multi channel images from a list of separate channel files. Expects results['img_filename'] to be a list of filenames. note that we r

An attention module used in BEVFormer based on Deformable-Detr. `Deformable DETR: Deformable Transformers for End-to-End Object Detection. <

An attention module used in BEVFormer based on Deformable-Detr. `Deformable DETR: Deformable Transformers for End-to-End Object Detection. <

Deconvolution upsample module in decoder for UNet (2X upsample). This module uses deconvolution to upsample feature map in the decoder of UNe

Interpolation upsample module in decoder for UNet. This module uses interpolation to upsample feature map in the decoder of UNet. It consists

UNet backbone. This backbone is the implementation of `U-Net: Convolutional Networks for Biomedical Image Segmentation <https://arxiv.org/abs

Base `TransformerLayer` for vision transformer. It can be built from `mmcv.ConfigDict` and support more flexible customization, for example,

This image file loader is for Waymo dataset. Load multi channel images from a list of separate channel files. Expects results['img_fi

Bbox coder for NMS-free detector. Args: pc_range (list[float]): Range of point cloud. post_center_range (list[float]): Limit of th

Normalize the image and add key "img_norm_cfg". Args: mean (sequence): Mean values of 3 channels. std (sequence): Std values

r"""NuScenes Dataset. This datset only add camera intrinsics and extrinsics to the results.

Implements decoder layer in DETR transformer. Args: attn_cfgs (list[`mmcv.ConfigDict`] | list[dict] | dict )): Configs for

Pad the multi-view image and add keys "pad_shape". There are two padding modes: (1) pad to a fixed size (2) pad to the minimum size that

Implements the Detr3D transformer. Args: as_two_stage (bool): Generate query from encoder features. Default: False.

Apply photometric distortion to image sequentially, every transformation is applied with a probability of 0.5. The position of random contrast is

Implements the Detr3D transformer. Args: as_two_stage (bool): Generate query from encoder features. Default: False.

Random scale the image Args: scales (list): List of scales to choose from. If the list contains only one scale

FPN used in SECOND/PointPillars/PartA2/MVXNet. Args: in_channels (list[int]): Input channels of multi-scale feature maps. out_cha

Head of Detr3D. Args: with_box_refine (bool): Whether to refine the reference points in the decoder. Defaults to False.