Matthew Muckley

Hello @fschiffers, that code is necessary for proper casting when using the `.to()` function of a module. Otherwise it may convert some tensors to real when they should remain complex.

@Xieyyyy in this case we need to enforce dependencies - the package doesn't work on PyTorch 1.11.

Great idea - do you have a specific example that might be useful for people executing 1D NUFFTs? If not, I could just throw something elementary together.

To me this sounds great @kylematoba - just don't have time to implement it right now myself due to other projects... If you open a PR would be happy to...

Hello @FredXL1, the oversampling should be cropped before the image is returned to you. I suspect the issue is with your k-space coordinates. Please be sure they're in the correct...

Hello @ajlok3, I'm not sure this is possible with more recent versions of PyTorch - as you can see [here](https://pytorch.org/docs/stable/generated/torch.fft.fft2.html#torch.fft.fft2) there are no options for real-valued tensors. If you pass...

Also, could you let me know what version of PyTorch you're on? As you can see [here](https://github.com/pytorch/pytorch/issues/45760) on more recent versions there might be complex-valued NCCL support.

Hello @kylematoba, in our case we define a forward NUFFT as a linear operator that applies an FFT followed by a linear interpolation with a Kaiser-Bessel function. Since it is...

@apt-get-nat using the duality property you may be able to approximate this operation using the package.

Hello @headmeister, you're correct that there is no planning stage. For the FFT step `torchkbnufft` puts all of that on the PyTorch FFT functions. Could you let me know what...