deepdream-cli-docker hangs in ubuntu 18.04

[bwanaaa]

I successfully installed docker community edition and it runs fine. Even added myself as a user so I dont have to type sudo before every docker command. I start with

cat /home/stefan/Pictures/DSC_1789.JPG | docker run -i saturnism/deepdream-cli > output.jpg

and get

libdc1394 error: Failed to initialize libdc1394
/usr/local/lib/python2.7/dist-packages/matplotlib/font_manager.py:273: UserWarning: Matplotlib is building the font cache using fc-list. This may take a moment.
  warnings.warn('Matplotlib is building the font cache using fc-list. This may take a moment.')
WARNING: Logging before InitGoogleLogging() is written to STDERR
W0323 17:52:28.385108     1 _caffe.cpp:122] DEPRECATION WARNING - deprecated use of Python interface
W0323 17:52:28.385124     1 _caffe.cpp:123] Use this instead (with the named "weights" parameter):
W0323 17:52:28.385126     1 _caffe.cpp:125] Net('tmp.prototxt', 1, weights='caffe/models/bvlc_googlenet/bvlc_googlenet.caffemodel')
I0323 17:52:28.388671     1 net.cpp:58] Initializing net from parameters: 
name: "GoogleNet"
force_backward: true

.
.
.
.
lots of stuf
.
.
.
.

I0323 17:52:28.438112     1 net.cpp:283] Network initialization done.
I0323 17:52:28.544561     1 upgrade_proto.cpp:52] Attempting to upgrade input file specified using deprecated V1LayerParameter: caffe/models/bvlc_googlenet/bvlc_googlenet.caffemodel
I0323 17:52:28.577404     1 upgrade_proto.cpp:60] Successfully upgraded file specified using deprecated V1LayerParameter
I0323 17:52:28.577956     1 net.cpp:761] Ignoring source layer label_data_1_split
I0323 17:52:28.578555     1 net.cpp:761] Ignoring source layer loss1/ave_pool
I0323 17:52:28.578560     1 net.cpp:761] Ignoring source layer loss1/conv
I0323 17:52:28.578562     1 net.cpp:761] Ignoring source layer loss1/relu_conv
I0323 17:52:28.578564     1 net.cpp:761] Ignoring source layer loss1/fc
I0323 17:52:28.578567     1 net.cpp:761] Ignoring source layer loss1/relu_fc
I0323 17:52:28.578567     1 net.cpp:761] Ignoring source layer loss1/drop_fc
I0323 17:52:28.578570     1 net.cpp:761] Ignoring source layer loss1/classifier
I0323 17:52:28.578572     1 net.cpp:761] Ignoring source layer loss1/loss
I0323 17:52:28.579334     1 net.cpp:761] Ignoring source layer loss2/ave_pool
I0323 17:52:28.579339     1 net.cpp:761] Ignoring source layer loss2/conv
I0323 17:52:28.579340     1 net.cpp:761] Ignoring source layer loss2/relu_conv
I0323 17:52:28.579342     1 net.cpp:761] Ignoring source layer loss2/fc
I0323 17:52:28.579344     1 net.cpp:761] Ignoring source layer loss2/relu_fc
I0323 17:52:28.579346     1 net.cpp:761] Ignoring source layer loss2/drop_fc
I0323 17:52:28.579349     1 net.cpp:761] Ignoring source layer loss2/classifier
I0323 17:52:28.579350     1 net.cpp:761] Ignoring source layer loss2/loss
I0323 17:52:28.581341     1 net.cpp:761] Ignoring source layer loss3/loss3
/usr/lib/python2.7/dist-packages/scipy/ndimage/interpolation.py:532: UserWarning: From scipy 0.13.0, the output shape of zoom() is calculated with round() instead of int() - for these inputs the size of the returned array has changed.
  "the returned array has changed.", UserWarning)

and then it just hangs there

htop shows this

https://imgur.com/yAkmA7D

Waiting 15' nothing happens

ctrl-C and I get

Traceback (most recent call last):
  File "./deepdream.py", line 135, in <module>
    out = deepdream(net, img, end=options.layer, iter_n=options.iterations, octave_n=options.octaves, octave_scale=options.scale)
  File "./deepdream.py", line 116, in deepdream
    make_step(net, end=end, clip=clip, **step_params)
  File "./deepdream.py", line 86, in make_step
    net.backward(start=end)
  File "/deepdream/caffe/distribute/python/caffe/pycaffe.py", line 169, in _Net_backward
    self._backward(start_ind, end_ind)
KeyboardInterrupt

Interestingly, when I run this

cat /home/stefan/Pictures/DSC_1789.JPG | docker run saturnism/deepdream-cli > output.jpg

(notice the absence of the -i flag) I get this error

Traceback (most recent call last):
  File "./deepdream.py", line 134, in <module>
    img = np.float32(PIL.Image.open(StringIO(sys.stdin.read())))
  File "/usr/lib/python2.7/dist-packages/PIL/Image.py", line 2028, in open
    raise IOError("cannot identify image file")
IOError: cannot identify image file

and when I run this

cat /home/stefan/Pictures/DSC_1789.JPG | docker run -iterations=1 saturnism/deepdream-cli > output.jpg

(notice -iterations=1 flag)

I get

the input device is not a TTY

running

docker run -i saturnism/deepdream-cli -h

gives

libdc1394 error: Failed to initialize libdc1394
/usr/local/lib/python2.7/dist-packages/matplotlib/font_manager.py:273: UserWarning: Matplotlib is building the font cache using fc-list. This may take a moment.
  warnings.warn('Matplotlib is building the font cache using fc-list. This may take a moment.')
Usage: deepdream.py [options]

Options:
  -h, --help            show this help message and exit
  -i ITERATIONS, --iterations=ITERATIONS
                        Iterations
  -o OCTAVES, --octaves=OCTAVES
                        Octatves
  -s SCALE, --scale=SCALE
                        Octatves Scale
  -l LAYER, --layer=LAYER
                        Layer
  --layers              Display layers

but none of the other flags (-o, -s, -l, --layers) work

[saturnism]

and when I run this

cat /home/stefan/Pictures/DSC_1789.JPG | docker run -iterations=1 saturnism/deepdream-cli > output.jpg

(notice -iterations=1 flag)

I get

the input device is not a TTY

This is because you put -iterations as a flag for docker run. The flag needs to be set for the container instance, docker run -i saturnism/deepdream-cli --iterations=1 ...

[saturnism]

htop shows this

https://imgur.com/yAkmA7D

Waiting 15' nothing happens

it looks like it's using 100% of 1 of the CPUs. The process does take some time. Can you try a smaller photo/picture?

[bwanaaa]

thank you for your replies. I am learning still about docker.

I resized the image to 400x300 pixels. The command executed to completion in 2 minutes. I thought I was in for another hang since the last lines I saw were:

/usr/lib/python2.7/dist-packages/scipy/ndimage/interpolation.py:532: UserWarning: From scipy 0.13.0, the output shape of zoom() is calculated with round() instead of int() - for these inputs the size of the returned array has changed.
  "the returned array has changed.", UserWarning)

but then I got the bash prompt back. whew.

I appreciate the work you put into the notebook as well. In it, you listed all the layers that the algorithm develops. I modified the script to loop through them all. Some interesting variations. However, some of the layers fail - specifically the ones with the word 'split' at the end

but then