cuboid_patch_based_geometric_compression

Bases: geometric_compression

The cuboid patch based geometric compression function.

It performs the data compression based on cuboid geometric patch shapes and provided compression metric. This class inherits from the geometric_compression class, and only redefines the initialization method.

...

Attributes:

Name	Type	Description
metric	Callable[[Tensor], Tensor]	The geometric compression metric.
name	str, default = 'cuboid_patch_based_geometric_compression'	Name of the compression function.
p_h	int	Height of the cuboid patch along the negative direction.
p_h_prime	int, default = None	Height of the cuboid patch along the positive direction.
p_w	int	Width of the cuboid patch along the negative direction.
p_w_prime	int, default = None	Width of the cuboid patch along the positive direction.
p_d	int	Depth of the cuboid patch along the negative direction.
p_d_prime	int, default = None	Depth of the cuboid patch along the positive direction.

Methods:

Name	Description
__init__	It performs the initialization of the compression function based on the provided metric and cuboid patch shapes.

Source code in tinybig/compression/geometric_compression.py

class cuboid_patch_based_geometric_compression(geometric_compression):
    """
        The cuboid patch based geometric compression function.

        It performs the data compression based on cuboid geometric patch shapes and provided compression metric.
        This class inherits from the geometric_compression class, and only redefines the initialization method.

        ...

        Attributes
        ----------
        metric: Callable[[torch.Tensor], torch.Tensor]
            The geometric compression metric.
        name: str, default = 'cuboid_patch_based_geometric_compression'
            Name of the compression function.
        p_h: int
            Height of the cuboid patch along the negative direction.
        p_h_prime: int, default = None
            Height of the cuboid patch along the positive direction.
        p_w: int
            Width of the cuboid patch along the negative direction.
        p_w_prime: int, default = None
            Width of the cuboid patch along the positive direction.
        p_d: int
            Depth of the cuboid patch along the negative direction.
        p_d_prime: int, default = None
            Depth of the cuboid patch along the positive direction.

        Methods
        ----------
        __init__
            It performs the initialization of the compression function based on the provided metric and cuboid patch shapes.
    """
    def __init__(
        self,
        metric: Callable[[torch.Tensor], torch.Tensor],
        p_h: int, p_w: int, p_d: int,
        p_h_prime: int = None, p_w_prime: int = None, p_d_prime: int = None,
        name: str = 'cuboid_patch_based_geometric_compression',
        *args, **kwargs
    ):
        """
            The initialization method of the cuboid patch based geometric compression function.

            It initializes the cuboid compression function based on the provided cuboid patch shape
            of the data segments to be compressed in the provided grid structure.

            Parameters
            ----------
            metric: Callable[[torch.Tensor], torch.Tensor]
                The geometric compression metric.
            name: str, default = 'cuboid_patch_based_geometric_compression'
                Name of the compression function.
            p_h: int
                Height of the cuboid patch along the negative direction.
            p_h_prime: int, default = None
                Height of the cuboid patch along the positive direction.
            p_w: int
                Width of the cuboid patch along the negative direction.
            p_w_prime: int, default = None
                Width of the cuboid patch along the positive direction.
            p_d: int
                Depth of the cuboid patch along the negative direction.
            p_d_prime: int, default = None
                Depth of the cuboid patch along the positive direction.

            Returns
            ----------
            transformation
                The cuboid geometric compression function.
        """
        patch = cuboid(p_h=p_h, p_w=p_w, p_d=p_d, p_h_prime=p_h_prime, p_w_prime=p_w_prime, p_d_prime=p_d_prime)
        super().__init__(name=name, metric=metric, patch=patch, *args, **kwargs)

__init__(metric, p_h, p_w, p_d, p_h_prime=None, p_w_prime=None, p_d_prime=None, name='cuboid_patch_based_geometric_compression', *args, **kwargs)

The initialization method of the cuboid patch based geometric compression function.

It initializes the cuboid compression function based on the provided cuboid patch shape of the data segments to be compressed in the provided grid structure.

Parameters:

Name	Type	Description	Default
metric	Callable[[Tensor], Tensor]	The geometric compression metric.	required
name	str	Name of the compression function.	'cuboid_patch_based_geometric_compression'
p_h	int	Height of the cuboid patch along the negative direction.	required
p_h_prime	int	Height of the cuboid patch along the positive direction.	None
p_w	int	Width of the cuboid patch along the negative direction.	required
p_w_prime	int	Width of the cuboid patch along the positive direction.	None
p_d	int	Depth of the cuboid patch along the negative direction.	required
p_d_prime	int	Depth of the cuboid patch along the positive direction.	None

Returns:

Type	Description
transformation	The cuboid geometric compression function.

Source code in tinybig/compression/geometric_compression.py

def __init__(
    self,
    metric: Callable[[torch.Tensor], torch.Tensor],
    p_h: int, p_w: int, p_d: int,
    p_h_prime: int = None, p_w_prime: int = None, p_d_prime: int = None,
    name: str = 'cuboid_patch_based_geometric_compression',
    *args, **kwargs
):
    """
        The initialization method of the cuboid patch based geometric compression function.

        It initializes the cuboid compression function based on the provided cuboid patch shape
        of the data segments to be compressed in the provided grid structure.

        Parameters
        ----------
        metric: Callable[[torch.Tensor], torch.Tensor]
            The geometric compression metric.
        name: str, default = 'cuboid_patch_based_geometric_compression'
            Name of the compression function.
        p_h: int
            Height of the cuboid patch along the negative direction.
        p_h_prime: int, default = None
            Height of the cuboid patch along the positive direction.
        p_w: int
            Width of the cuboid patch along the negative direction.
        p_w_prime: int, default = None
            Width of the cuboid patch along the positive direction.
        p_d: int
            Depth of the cuboid patch along the negative direction.
        p_d_prime: int, default = None
            Depth of the cuboid patch along the positive direction.

        Returns
        ----------
        transformation
            The cuboid geometric compression function.
    """
    patch = cuboid(p_h=p_h, p_w=p_w, p_d=p_d, p_h_prime=p_h_prime, p_w_prime=p_w_prime, p_d_prime=p_d_prime)
    super().__init__(name=name, metric=metric, patch=patch, *args, **kwargs)