Loading...
Close
XCurveXCurve
Sidebar

Search tutorials

Get Started

Documentation

Example Projects

XCurve.Losses.AUROC

The StandardAUROC Loss consists of three parts: SquareAUCLoss, HingeAUCLoss, and ExpAUCLoss.

SquareAUCLoss

The following class supports the computation of SquareAUCloss: 
  CLASS SquareAUCLoss (num_classes, gamma=1, transform='ovo', **kwargs)     [SOURCE]
Parameters:
num_classes (int)

The number of classes in the classification problem.

gamma (float)

The safe margin for the loss (default=1.0).

transform (str)

The manner of computing the multi-class AUROC Metric, either 'ovo' or 'ova' (default as 'ovo' in our paper).

Example: 
  import torch
  from XCurve.AUROC.losses import SquareAUCLoss
                
  feats = torch.softmax(torch.randn(16, 9).cuda(), dim=1)
  targets = torch.tensor([0,1,2,3,4,5,6,7,8,7,6,5,4,3,2,1]).cuda()
  criterion_Square = SquareAUCLoss(
  num_classes=9,
  gamma=1.0,
  transform="ovo")
  loss_Square = criterion_Square(feats, targets)
  print(loss_Square.item())

HingeAUCLoss

The following class supports the computation of HingeAUCloss: 
  CLASS HingeAUCLoss (num_classes, gamma=1, transform='ovo', **kwargs)     [SOURCE]
Parameters:
num_classes (int)

The number of classes in the classification problem.

gamma (float)

The safe margin for the loss (default=1.0).

transform (str)

The manner of computing the multi-class AUROC Metric, either 'ovo' or 'ova' (default as 'ovo' in our paper).

Example: 
  import torch
  from XCurve.AUROC.losses import HingeAUCLoss
                
  feats = torch.softmax(torch.randn(16, 9).cuda(), dim=1)
  targets = torch.tensor([0,1,2,3,4,5,6,7,8,7,6,5,4,3,2,1]).cuda()
  criterion_Hinge = HingeAUCLoss(
  num_classes=9,
  gamma=1.0,
  transform="ovo")
  loss_Hinge = criterion_Hinge(feats, targets)
  print(loss_Hinge.item())

ExpAUCLoss

The following class supports the computation of ExpAUCloss: 
  CLASS ExpAUCLoss (num_classes, gamma=1, transform='ovo', **kwargs)     [SOURCE]
Parameters:
num_classes (int)

The number of classes in the classification problem.

gamma (float)

The safe margin for the loss (default=1.0).

transform (str)

The manner of computing the multi-class AUROC Metric, either 'ovo' or 'ova' (default as 'ovo' in our paper).

Example: 
  import torch
  from XCurve.AUROC.losses import ExpAUCLoss
                
  feats = torch.softmax(torch.randn(16, 9).cuda(), dim=1)
  targets = torch.tensor([0,1,2,3,4,5,6,7,8,7,6,5,4,3,2,1]).cuda()
  criterion_Exp = ExpAUCLoss(
    num_classes=9,
    gamma=1.0,
    transform="ovo")
  loss_Exp = criterion_Exp(feats, targets)
  print(loss_Exp.item())