Problem in data-consistency module

[bilalkabas]

Hi,

Thank you for this great work. While using the repo, I came accross with a problem in data-consistency module. Below is a simple forward pass to the data-consistency module. The problem is in k-space transformation. I wanted to contribute to the repo with a PR addressing this problem. It may help those working on this repo.

import torch
from backbones.reconformer.reconformer import DataConsistencyInKspace

resolution = 320
device = 'cuda:0'

x = torch.randn((1, 2, resolution, resolution)).to(device)
k0 = torch.randn((1, 2, resolution, resolution)).to(device)
mask = torch.randn((1, 1, resolution, resolution)).to(device)

dc = DataConsistencyInKspace()
out = dc(x, k0, mask)

print(f"Input shape: {x.shape}")
print(f"Output shape: {out.shape}")

     46 k0 = k0.permute(0, 2, 3, 1)
     47 mask = mask.permute(0, 2, 3, 1)
...
--> 122 data = torch.fft.fft(data, 2, normalized=normalized)
    123 data = fftshift(data, dim=(-3, -2))
    124 return data

TypeError: fft_fft() got an unexpected keyword argument 'normalized'