Custom exception for catching expected errors.

Visdom Line Visualization Helper Class.

This is the Deformable DETR module that performs object detection

Very simple multi-layer perceptron (also called FFN)

Simple convolutional head, using group norm. Upsampling is done using a FPN approach

Track a series of values and provide access to smoothed values over a window or the global series average.

Swin Transformer backbone. A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` - https:/

ResNet backbone with frozen BatchNorm.

A basic Swin Transformer layer for one stage. Args: dim (int): Number of feature channels depth (int): Depths of this stage.

Class which simply counts the number of tracker and gt detections and ids.

This class computes an assignment between the targets and the predictions of the network For efficiency reasons, the targets don't include the no

This class computes an assignment between the targets and the predictions of the network For efficiency reasons, the targets don't include the no

Multilayer perceptron.

Image to Patch Embedding Args: patch_size (int): Patch token size. Default: 4. in_chans (int): Number of input image channels. De

Standard 2D sine positional encoding

Extends 2D x,y sine positional encoding with learned temporal embedding

Adapted from VisTr, positional encoding for x,y,t simultaneously. Pads with 0's if hidden size is not divisible by 3.

This class computes the loss for DeVIS. The process happens in two steps: 1) we compute hungarian assignment between ground truth boxes a

ResNet backbone with frozen BatchNorm.

Swin Transformer Block. Args: dim (int): Number of input channels. num_heads (int): Number of attention heads. window_siz

Window based multi-head self attention (W-MSA) module with relative position bias. It supports both of shifted and non-shifted window. Args:

Dataset class for BDD100K tracking

Class which implements the CLEAR metrics

Dataset class for DAVIS tracking

Evaluator class for evaluating different metrics for different datasets

BatchNorm2d where the batch statistics and the affine parameters are fixed. Copy-paste from torchvision.misc.ops with added eps before rqsrt

Class which implements the HOTA metrics. See: https://link.springer.com/article/10.1007/s11263-020-01375-2

Dataset class for Head Tracking Challenge - 2D bounding box tracking

Class which implements the ID metrics

Class which implements the ID metrics

Visdom Image Visualization Helper Class.

Class which implements the J&F metrics

Dataset class for KITTI 2D bounding box tracking

Dataset class for KITTI MOTS tracking

Dataset class for MOTS Challenge tracking

Dataset class for MOT Challenge 2D bounding box tracking

Patch Merging Layer Args: dim (int): Number of input channels. norm_layer (nn.Module, optional): Normalization layer. Default: n