bartorch.fft

bartorch.fft(input, axes, *, inverse=False, unitary=False, uncentred=False)

Fourier transform along axes, centred by default.

Parameters:

• input (torch.Tensor)

• axes (int or tuple of int) – Axes to transform, as indices into input.shape.

• inverse (bool) – Transform with the positive exponent (-i).

• unitary (bool) – Scale by one over the square root of the transformed size (-u); otherwise unnormalized.

• uncentred (bool) – Keep the zero frequency at index zero rather than at n // 2 (-n).

Examples

>>> spectrum = fft(image, axes=(-2, -1), unitary=True)