vagg19_net KeyError normalization at matconvnet.py

[LeeHyungDo]

i am trying to run this code. and i completed setting the requirement( Deeppy, cuDnn, choose imagenet-vgg-verydeep-19), vgg-verydeep-19 file is located at this project directory.

but i get error like below.

Traceback (most recent call last):
File "neural_artistic_style.py", line 139, in
run()
File "neural_artistic_style.py", line 99, in run
layers, img_mean = vgg19_net(args.vgg19, pool_method=args.pool_method)
File "/home/ubuntu/LeeHD/neural_artistic_style-master/neural_artistic_style-master/matconvnet.py", line 35, in vgg19_net
img_mean = matconvnet['normalization'][0][0][0].astype(dp.float_)
KeyError: 'normalization'

what should i do to fix this error?

[rnikhill]

MatConvNet recently switched over from beta16 to beta17 and changed their pre-trained models accordingly. Neural-Artistic-Style apparently only works with the beta16 format, which you can download from here

[andersbll]

Hey, thank you both for the info. I have now switched to the new file format. Hopefully it won't change anytime soon. :)

[kapyar]

Hey, I got the same issue and when I download new imagenet-vgg-verydeep-19.mat from beta16 then I got following problem

CUDArray: CUDA back-end not available, using NumPy. Traceback (most recent call last): File "neural_artistic_style.py", line 139, in <module> run() File "neural_artistic_style.py", line 117, in run args.smoothness) File "/Users/yaroslav/Documents/NaUKMA/NeuronArtist/neural_artistic_style/style_network.py", line 88, in __init__ self.x._setup(x_shape) AttributeError: 'Parameter' object has no attribute '_setup'

Does someone know how to solve this problem?
Thanks

[andersbll]

Hi, I think you need to pull the latest version of DeepPy/Neural artistic style.

[kapyar]

Did you mean this one https://github.com/andersbll/deeppy ?

[andersbll]

Yes.

[kapyar]

Thanks, I've sold this problem, I think the reason was that I wasn't use sudo command when run python setup.py install