Complex number DataLoader

Hi, I’m new to Pytorch.
I want to train Autoencoder and Denoising Autoencoders with my 30k complex number array data.
I saved every single array(1,480,480) as npy.
How should I make a data loader for AE and DAE with this data?
I mean, I tried to put it into one list (yeah list with 30k elements) make a tensor from this, it failed.
My complex number arrays represent images after FFT.
I will be grateful for any clue how to work with this data in PyTorch.

What exactly failed while trying to create the tensor? Could you post the error message and the corresponding code, please?

OK, I should use torch.from_numpy instead torch.ComplexType.
Now It works.