torchaudio.functional.resample() doesn't work with complex types

[pfeatherstone]

🐛 Describe the bug

torchaudio.functional.resample() doesn't work with complex types.

Example:

x = torch.randn(1024, 2)
x = torch.view_as_complex(x)
y = torchaudio.functional.resample(x, 1.0, 2.0)

Gives error:

RuntimeError: "arange_cpu" not implemented for 'ComplexFloat'

Versions

>>> import torch
>>> import torchaudio
>>> torch.__version__
'2.0.0+cu117'
>>> torchaudio.__version__
'2.0.1+cu117'

[hwangjeff]

Hi @pfeatherstone, thank you for opening the issue. Could you share more details on your use case for us to figure out to how to address this?

[pfeatherstone]

I wanted to resample some IQ data. It's not a problem. The filter is real-valued so you can resample real and imaginary separately.