Pooling

class complextorch.nn.modules.pooling.CVAdaptiveAvgPool1d(output_size: int | Tuple[int])

1-D Complex-Valued Adaptive Average Pooling

Applies adaptive average pooling using torch.nn.AdaptiveAvgPool1d to the real and imaginary parts of the input tensor separately.

Implements the following operation:

\[G(\mathbf{z}) = \texttt{AdaptiveAvgPool1d}(\mathbf{x}) + j \texttt{AdaptiveAvgPool1d}(\mathbf{y}),\]

where \(\mathbf{z} = \mathbf{x} + j\mathbf{y}\)

forward(input: CVTensor) CVTensor

Applies adaptive average pooling using torch.nn.AdaptiveAvgPool1d to the real and imaginary parts of the input tensor separately.

Parameters:

input (CVTensor) – input tensor

Returns:

\(\texttt{AdaptiveAvgPool1d}(\mathbf{x}) + j \texttt{AdaptiveAvgPool1d}(\mathbf{y})\)

Return type:

CVTensor

class complextorch.nn.modules.pooling.CVAdaptiveAvgPool2d(output_size)

2-D Complex-Valued Adaptive Average Pooling

Applies adaptive average pooling using torch.nn.AdaptiveAvgPool2d to the real and imaginary parts of the input tensor separately.

Implements the following operation:

\[G(\mathbf{z}) = \texttt{AdaptiveAvgPool2d}(\mathbf{x}) + j \texttt{AdaptiveAvgPool2d}(\mathbf{y}),\]

where \(\mathbf{z} = \mathbf{x} + j\mathbf{y}\)

forward(input: CVTensor) CVTensor

Applies adaptive average pooling using torch.nn.AdaptiveAvgPool2d to the real and imaginary parts of the input tensor separately.

Parameters:

input (CVTensor) – input tensor

Returns:

\(\texttt{AdaptiveAvgPool2d}(\mathbf{x}) + j \texttt{AdaptiveAvgPool2d}(\mathbf{y})\)

Return type:

CVTensor

class complextorch.nn.modules.pooling.CVAdaptiveAvgPool3d(output_size)

3-D Complex-Valued Adaptive Average Pooling

Applies adaptive average pooling using torch.nn.AdaptiveAvgPool3d to the real and imaginary parts of the input tensor separately.

Implements the following operation:

\[G(\mathbf{z}) = \texttt{AdaptiveAvgPool3d}(\mathbf{x}) + j \texttt{AdaptiveAvgPool3d}(\mathbf{y}),\]

where \(\mathbf{z} = \mathbf{x} + j\mathbf{y}\)

forward(input: CVTensor) CVTensor

Applies adaptive average pooling using torch.nn.AdaptiveAvgPool3d to the real and imaginary parts of the input tensor separately.

Parameters:

input (CVTensor) – input tensor

Returns:

\(\texttt{AdaptiveAvgPool3d}(\mathbf{x}) + j \texttt{AdaptiveAvgPool3d}(\mathbf{y})\)

Return type:

CVTensor