yangyan92 / admm-csnet Goto Github PK

When I run the testing code in "Generic-ADMM-CSNet-RealMRI", I got the following error message:

Attempt to execute SCRIPT vl_nnconv as a function:
.\ADMM-Net_package\matlab\vl_nnconv.m

Error in before_yloss_with_gradient_single (line 56)
res(i+1).x = vl_nnconv(res(i).x, w1, w2, ...

My guess is that the mex files are not compiled in the same environment. Any idea how to resolve this? I am running on Matlab 2018a, Windows version.

Dear Yang Yan,

In your provided data link, I found no training data of ComplexMRI. Could you please offer me the training data? I am going to learn your paper and code.

Best wishes,
Gao

When running each line of main_ADMM_CSNet_test.m in CSNet-Image, I get an error on the following line:

[re_LOss, re_PSnr , re_SSim, rec_image ]= loss_with_gradient_single_before(x0, im_ori, net, mask.picks, mask.perm);

Attempt to execute SCRIPT vl_nnconv as a function:
C:\Users\Oliver\Desktop\ADMM-CSNet-master\matlab\vl_nnconv.m
Error in dagnn.Conv2D/forward (line 10)
outputs{1} = vl_nnconv(inputs{1}, params{1}, params{2}, 'pad', pad1);
Error in dagnn.Layer/forwardAdvanced (line 85)
outputs = obj.forward(inputs, {net.params(par).value}) ;
Error in dagnn.DagNN/eval (line 91)
obj.layers(l).block.forwardAdvanced(obj.layers(l)) ;
Error in loss_with_gradient_single_before (line 41)
net.eval(inputs) ;

Is there a particular way to resolve this to test the code? I'm running this on MATLAB R2018a.