espnet2.enh.layers.complex_utils.inverse

source

espnet2.enh.layers.complex_utils.inverse(c: Tensor | ComplexTensor) → Tensor | ComplexTensor

Compute the inverse of a complex or real tensor.

This function calculates the inverse of a given tensor, which can be either a real tensor or a complex tensor (using ComplexTensor). If the input is a ComplexTensor, the function will use the inverse2() method. For real tensors, it will use the standard inverse() method.

• Parameters:c – A tensor that can either be a torch.Tensor or a ComplexTensor. It is expected to be a square matrix for the inverse operation.
• Returns: A tensor that is the inverse of the input tensor c. The type of the returned tensor will match the input tensor type.
• Raises:ValueError – If the input tensor is not square or of unsupported type.

Examples

>>> import torch
>>> from torch_complex.tensor import ComplexTensor
>>> real_tensor = torch.tensor([[1.0, 2.0], [3.0, 4.0]])
>>> complex_tensor = ComplexTensor(torch.tensor([[1.0, 2.0], [3.0, 4.0]]),
...                                torch.tensor([[0.0, 0.0], [0.0, 0.0]]))
>>> inverse_real = inverse(real_tensor)
>>> inverse_complex = inverse(complex_tensor)
>>> print(inverse_real)
tensor([[-2.0000,  1.0000],
        [1.5000, -0.5000]])
>>> print(inverse_complex)
ComplexTensor(real=tensor([[-2.0000,  1.0000],
                            [1.5000, -0.5000]]),
              imag=tensor([[0.0, 0.0],
                            [0.0, 0.0]]))

NOTE

The input tensor must be a square matrix for the inverse to be computed. Ensure that the dimensions of the input tensor are valid to avoid runtime errors.