Skip to content

swav_loss

SwaVLoss

Bases: SwaVLoss

A class representing a custom SwaV loss function.

Attributes:

Name Type Description
temperature float

The temperature parameter for the loss calculation. Default is 0.1.
sinkhorn_iterations int

The number of iterations for Sinkhorn algorithm. Default is 3.
sinkhorn_epsilon float

The epsilon parameter for Sinkhorn algorithm. Default is 0.05.
sinkhorn_gather_distributed bool

Whether to gather distributed results for Sinkhorn algorithm. Default is False.
Source code in fmcib/ssl/losses/swav_loss.py 
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
class SwaVLoss(lightly.loss.swav_loss.SwaVLoss):
    """
    A class representing a custom SwaV loss function.

    Attributes:
        temperature (float): The temperature parameter for the loss calculation. Default is 0.1.
        sinkhorn_iterations (int): The number of iterations for Sinkhorn algorithm. Default is 3.
        sinkhorn_epsilon (float): The epsilon parameter for Sinkhorn algorithm. Default is 0.05.
        sinkhorn_gather_distributed (bool): Whether to gather distributed results for Sinkhorn algorithm. Default is False.
    """

    def __init__(
        self,
        temperature: float = 0.1,
        sinkhorn_iterations: int = 3,
        sinkhorn_epsilon: float = 0.05,
        sinkhorn_gather_distributed: bool = False,
    ):
        """
        Initialize the object with specified parameters.

        Args:
            temperature (float, optional): The temperature parameter. Default is 0.1.
            sinkhorn_iterations (int, optional): The number of Sinkhorn iterations. Default is 3.
            sinkhorn_epsilon (float, optional): The epsilon parameter for Sinkhorn algorithm. Default is 0.05.
            sinkhorn_gather_distributed (bool, optional): Whether to use distributed computation for Sinkhorn algorithm. Default is False.
        """
        super().__init__(temperature, sinkhorn_iterations, sinkhorn_epsilon, sinkhorn_gather_distributed)

    def forward(self, pred):
        """
        Perform a forward pass of the model.

        Args:
            pred (tuple): A tuple containing the predicted outputs for high resolution, low resolution, and queue.

        Returns:
            The output of the forward pass.
        """
        high_resolution_outputs, low_resolution_outputs, queue_outputs = pred
        return super().forward(high_resolution_outputs, low_resolution_outputs, queue_outputs)

__init__(temperature=0.1, sinkhorn_iterations=3, sinkhorn_epsilon=0.05, sinkhorn_gather_distributed=False)

Initialize the object with specified parameters.

Parameters:

Name Type Description Default
temperature float

The temperature parameter. Default is 0.1.

 0.1
sinkhorn_iterations int

The number of Sinkhorn iterations. Default is 3.

 3
sinkhorn_epsilon float

The epsilon parameter for Sinkhorn algorithm. Default is 0.05.

 0.05
sinkhorn_gather_distributed bool

Whether to use distributed computation for Sinkhorn algorithm. Default is False.

 False
Source code in fmcib/ssl/losses/swav_loss.py 
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
def __init__(
    self,
    temperature: float = 0.1,
    sinkhorn_iterations: int = 3,
    sinkhorn_epsilon: float = 0.05,
    sinkhorn_gather_distributed: bool = False,
):
    """
    Initialize the object with specified parameters.

    Args:
        temperature (float, optional): The temperature parameter. Default is 0.1.
        sinkhorn_iterations (int, optional): The number of Sinkhorn iterations. Default is 3.
        sinkhorn_epsilon (float, optional): The epsilon parameter for Sinkhorn algorithm. Default is 0.05.
        sinkhorn_gather_distributed (bool, optional): Whether to use distributed computation for Sinkhorn algorithm. Default is False.
    """
    super().__init__(temperature, sinkhorn_iterations, sinkhorn_epsilon, sinkhorn_gather_distributed)

forward(pred)

Perform a forward pass of the model.

Parameters:

Name Type Description Default
pred tuple

A tuple containing the predicted outputs for high resolution, low resolution, and queue.

 required

Returns:

Type Description

The output of the forward pass.
Source code in fmcib/ssl/losses/swav_loss.py 
33
34
35
36
37
38
39
40
41
42
43
44
def forward(self, pred):
    """
    Perform a forward pass of the model.

    Args:
        pred (tuple): A tuple containing the predicted outputs for high resolution, low resolution, and queue.

    Returns:
        The output of the forward pass.
    """
    high_resolution_outputs, low_resolution_outputs, queue_outputs = pred
    return super().forward(high_resolution_outputs, low_resolution_outputs, queue_outputs)