Configurable identity-mapping residual block Parameters ---------- in_channels : int Number of input channels.

Creates an InPlace Activated Batch Normalization module Parameters ---------- num_features : int Number of featur

Computes and stores the Metric values from Confusion Matrix - overall accuracy - mean accuracy - mean IU

Bootstrapped CrossEntropy2D: The pixel-bootstrapped cross entropy loss :param weight: <torch.Tensor, optional> A manual rescaling we

Bootstrapped CrossEntropy2D: The pixel-bootstrapped cross entropy loss :param weight: <torch.Tensor, optional> A manual rescaling we

Loss(x, class) = - \alpha (1-softmax(x)[class])^gamma \log(softmax(x)[class]) :param ignore_label: <int> ignore label :para

Semantic Encoding Loss :param num_classes: <int> Number of classes :param ignore_label: <int, optional> Specifies a target va

Soft-Jaccard Loss for Semantic Segmentation :param weight: <torch.Tensor, optional> A manual rescaling weight given to each class.

Tversky Loss for Semantic Segmentation :param alpha: <int> Parameter to control precision and recall :param beta: <int> Para

Tversky Loss for Semantic Segmentation :param beta: <int> Parameter to control precision and recall :param weight: <torch.Te

:param limit: <tuple of float> for example: (0.725, 1.25)

MobileNetV2Plus: MobileNetV2 based Semantic Segmentation :param num_classes: (int) Number of classes :param width_multi:

__len__

Cross GPU all reduce autograd operation for calculate mean and variance in SyncBN.

Add Cartesian Coordination Info to Current Tensor :param x: shape(N, C, H, W) :return: shape(N, C+2 or C+3, H, W)