Getting an error when trying to run

[aaronjolson]

All of the dependencies seemed to install correctly with pip. Any idea what might fix this?

C:\Users\aaols\Documents\large_neural_style>python large_neural_style.py --content eagle.jpg --style lock.jpg --output leagle.jpg

WARNING (theano.tensor.blas): Using NumPy C-API based implementation for BLAS functions.
Content image shape:  (2426, 2066, 3)
Pyramid sizes: [(240, 204, 3), (480, 408, 3), (960, 817, 3), (1920, 1635, 3)]
shape: (240, 204, 3)
C:\Python36\lib\site-packages\skimage\transform\_warps.py:105: UserWarning: The default mode, 'constant', will be changed to 'reflect' in skimage 0.15.
  warn("The default mode, 'constant', will be changed to 'reflect' in "
C:\Python36\lib\site-packages\skimage\transform\_warps.py:110: UserWarning: Anti-aliasing will be enabled by default in skimage 0.15 to avoid aliasing artifacts when down-sampling images.
  warn("Anti-aliasing will be enabled by default in skimage 0.15 to "
resized_content_image.shape:  (304, 268, 3)
content.shape: (1, 3, 304, 268)
style.shape: (1, 3, 152, 240)
after image_to_shape(), style.shape: (1, 3, 240, 204)
Traceback (most recent call last):
  File "large_neural_style.py", line 193, in <module>
    sys.exit(main())
  File "large_neural_style.py", line 129, in main
    args.print_loss)
  File "C:\Users\aaols\Documents\large_neural_style\stylize.py", line 30, in generate_stylized_image
    stylize = Stylize(content, True, style, content_weight, style_weight, style_layer_weight_factor, total_variation_weight, print_loss, print_progress=True)
  File "C:\Users\aaols\Documents\large_neural_style\stylize.py", line 92, in __init__
    self.net.load_params('vgg19_normalized.pkl')
  File "C:\Users\aaols\Documents\large_neural_style\vgg19.py", line 37, in load_params
    values = pickle.load(open(weights_filename))['param values']
  File "C:\Python36\lib\encodings\cp1252.py", line 23, in decode
    return codecs.charmap_decode(input,self.errors,decoding_table)[0]
UnicodeDecodeError: 'charmap' codec can't decode byte 0x81 in position 3634: character maps to <undefined>

[aaronjolson]

I tried changing https://github.com/dnuffer/large_neural_style/blob/master/vgg19.py#L37 to values = pickle.load(open(weights_filename, encoding="utf8"))['param values'] and I get a slightly different error

Traceback (most recent call last):
  File "large_neural_style.py", line 193, in <module>
    sys.exit(main())
  File "large_neural_style.py", line 129, in main
    args.print_loss)
  File "C:\Users\aaols\Documents\large_neural_style\stylize.py", line 30, in generate_stylized_image
    stylize = Stylize(content, True, style, content_weight, style_weight, style_layer_weight_factor, total_variation_weight, print_loss, print_progress=True)
  File "C:\Users\aaols\Documents\large_neural_style\stylize.py", line 92, in __init__
    self.net.load_params('vgg19_normalized.pkl')
  File "C:\Users\aaols\Documents\large_neural_style\vgg19.py", line 37, in load_params
    values = pickle.load(open(weights_filename, encoding="utf8"))['param values']
  File "C:\Python36\lib\codecs.py", line 321, in decode
    (result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte

So I am assuming this is some kind of encoding error.

[dnuffer]

The error is because the weights were pickled with python 2.7 and you're using 3.6. Apparently there's potential incompatibilities. Reading https://stackoverflow.com/questions/28218466/unpickling-a-python-2-object-with-python-3 it seems you can try passing encoding="latin1" to pickle.load(), and if that doesn't work, I'd suggest using python 2.7.

[dnuffer]

Also you can try using encoding='bytes'

[aaronjolson]

Ah, I see, thank you!
I changed line 37 to values = pickle.load(open(weights_filename, 'rb'), encoding='latin-1')['param values'] and that has moved me forward.