Hi,
I have seen that the function torch.fft.rfft only applies the real fast Fourier transform across the last dimension of your input, and applies the usual fast Fourier transform across of the rest of the dimensions.
Is there a way to apply rfft across all the dimensions of your input?