RuntimeError: "max_cpu" not implemented for 'ComplexFloat'

1

Hello,
I have a tensor of shape torch.Size([128, 1, 480]) and I want to get the maximum value of the tensor. The tensor has float complex numbers and I am running my code on CPU, but the code results in the following error:
RuntimeError: “max_cpu” not implemented for ‘ComplexFloat’
It seems that the max function is not implemented for running on CPU yet. Is there any alternatives to use instead?
Thanks in advance.in

2

You’ll need to define what you mean by max value for complex numbers. You have a few options:

  1. Max real value;
  2. Max imaginary value;
  3. Max absolute value or magnitude;
  4. Max complex angle.