TCCE Class Reference

Inheritance diagram for TCCE:

Public Member Functions
def	__init__ (self, int n=0)
torch.tensor	forward (self, torch.tensor y, torch.LongTensor y_true)

Public Attributes
	n
	Number of Taylor terms.

Detailed Description

Class for calculation of Taylor cross entropy loss.

Attributes:
    n (int): Number of Taylor series terms to be used for loss calculation.

Definition at line 45 of file torch_tcce.py.

Constructor & Destructor Documentation

__init__()

def __init__	(		self,
		int	n = 0
	)

Initialize the loss class.

Parameters:
    n (int)(optional): Number of Taylor series terms to be used for loss calculation.

Definition at line 54 of file torch_tcce.py.

    def __init__(self, n: int = 0):
        """
        Initialize the loss class.
 
        Parameters:
            n (int)(optional): Number of Taylor series terms to be used for loss calculation.
 
        """
        super().__init__()
        
        self.n = n
 

Member Function Documentation

forward()

torch.tensor forward	(		self,
		torch.tensor	y,
		torch.LongTensor	y_true
	)

Calculates the Taylor categorical cross entropy loss.

Parameters:
    y(torch.tensor): Tensor containing the output of the model.
    y_true(torch.tensor): 1D tensor containing the truth value for a given set of features.

Returns:
    The calculated loss as a torch tensor.

Definition at line 66 of file torch_tcce.py.

    def forward(self, y: torch.tensor, y_true: torch.LongTensor) -> torch.tensor:
        """
        Calculates the Taylor categorical cross entropy loss.
 
        Parameters:
            y(torch.tensor): Tensor containing the output of the model.
            y_true(torch.tensor): 1D tensor containing the truth value for a given set of features.
 
        Returns:
            The calculated loss as a torch tensor.
        """
        return tcc(y, y_true, self.n)

Member Data Documentation

n

n

Number of Taylor terms.

Definition at line 64 of file torch_tcce.py.

---

The documentation for this class was generated from the following file:

• mva/scripts/basf2_mva_extensions/torch_tcce.py