visionai/clouddream

Deepdreaming in the clouds: A Dockerized deepdream Guide

clouddream's Issues

changing image size

Possible to change the output size of the processed images to larger?

My fork might be of interest

Hi,

Sent you an email a few weeks ago as I started working on my fork, but not sure you received it.
I've heavily changed your original code for my needs, and I thought it might be worth discussing how we could work together to make more people benefit from them.

My fork is at https://github.com/hamstah/clouddream

Some of the changes I introduced

• There is a web based manager that allows the uploading of new images by url/file upload
• You can browse the input/output of the file from there
• You can process any of the input images with parameters specified by the user (maxwidth, n_inter, n_octave etc), so you can have multiple outputs from the same image with different parameters
• Processing is ordered and done by workers, you can see the queue and jobs in progress (you can have multiple workers in parallel). I had issues with the original project where some image would fail and it would get stuck in a loop trying to process them.
• Use of docker compose to make it easier to spin the containers and scale the number of workers

It's obviously work in progress and I'm no docker expert, but I think it provides a good foundation to extend.

I'm thinking of moving the project out of the fork and into its own repo, as it's almost a complete rewrite.

Would be great to hear your thoughts.

I have a version of it running on a digital ocean droplet, I can give you access if you want to check it out.

Is it possible to do guided dreams in this build ?

Aside, the installation was really brain dead! I got it setup in 10 minutes and was able to produce my images. Thanks a lot for putting together such a clean setup. I would also love to train with my own dataset. So it would be awesome if you could support it someday.

Add a better way of pushing images to the server.

Hello.
I'm finding very difficult to push images to the server.
Can you add like a web file browser or a ftp server?(with password)
Also is possible to do something with the youtube urls?
Like in a text file we put youtube urls and the server process them.
Eg: youtube_url layertype

GPU support?

Any interest in adding it? Could open up some very cool realtime (or near-realtime) options.

Here's a preview of the video I'm working on. Each frame takes a few minutes to render on my DigitalOcean Droplet with 2GB Ram.

Make the GUI not say "Nice Photo, eh" for every image

The "Nice Photo" line can either go away, be randomly generated from a list of statements like "Cool photo" "Check this out" "Interesting", or mention the ConvNet layer targeted for the visualization like "conv2/3x3" (but this will require the generated json files to include this information).

Youtube Video Processing Does Not Work?

Whenever I launch the ./youtube.sh .... this is the output I get:
Could find no file with path '/tmp/outputs/image-%05d.jpg' and index in the range 0-4
/tmp/outputs/image-%05d.jpg: No such file or directory.

Anyone else get this?

Conversion failed - boot2docker + youtube.sh

@quantombone Thanks for putting together the docs on getting boot2 running, you're awesome.

I tried running the youtube script and get the following:

Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> mjpeg (native))
Press [q] to stop, [?] for help
[image2 @ 0x2cb4d20] Could not open file : /tmp/images/image-00001.jpg
av_interleaved_write_frame(): Input/output error
frame=    1 fps=0.0 q=5.2 Lsize=N/A time=00:00:00.03 bitrate=N/A    
video:12kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
Conversion failed!

I checked my tmp and both folders exist.

Adding models to the docker container

I'm attempting to add a new model to the docker container, but I'm missing something as I can't seem to get it to save.

I use ./enter.sh to enter the container, then run something like;

sudo apt-get install vim

exit the docker and run;

sudo docker commit 89028ee7020d visionai/clouddream

I then restart the containers, however my changes don't seem to be reflected in the now running container. Could you possibly let me know what step I'm missing here or where I'm going wrong?

Many thanks!

DigitalOcean API + Deepdream Example

There should be an example that starts with a digital ocean API token, and then does these steps:

1. Scaffold a new cloud instance (an Ubuntu appliance with Docker) "the server"
2. Checkout this repository on the instance on the server
3. Copy a directory of images from the client to the server
4. Perform all of the computation in the cloud and copy the results back to the client
5. Clean up the Application so that no more money has to be spent on keeping the DigitalOcean instance running

Сlouddream on centos or other Linux ?

Will work clouddream on centos? Docker not work in this OS, so all applications need to be installed manually

when try to stop get an error:

When I run the command
./stop.sh

I get this error:

Error response from daemon: No such container: deepdream-enter

Errors running youtube.sh

root@dreamzzz:~/clouddream# ./youtube.sh https://www.youtube.com/watch?v=nuHfVn_cfHU
URL is https://www.youtube.com/watch?v=nuHfVn_cfHU
[youtube] nuHfVn_cfHU: Downloading webpage
[youtube] nuHfVn_cfHU: Extracting video information
WARNING: [youtube] nuHfVn_cfHU: Skipping DASH manifest: ExtractorError(u'Cannot decrypt signature without player_url; please report this issue on https://yt-dl.org/bug . Make sure you are using the latest version; see  https://yt-dl.org/update  on how to update. Be sure to call youtube-dl with the --verbose flag and include its complete output.',)
[download] Destination: /tmp/video.mp4
[download] 100% of 9.40MiB in 00:03
ffmpeg version git-2014-07-28-a776238 Copyright (c) 2000-2014 the FFmpeg developers
  built on Jul 28 2014 03:08:49 with gcc 4.8 (Ubuntu 4.8.2-19ubuntu1)
  configuration: --shlibdir=/usr/lib64 --prefix=/usr --mandir=/usr/share/man --libdir=/usr/lib64 --enable-static --extra-cflags='-fmessage-length=0 -grecord-gcc-switches -fstack-protector -O2 -Wall -D_FORTIFY_SOURCE=2 -funwind-tables -fasynchronous-unwind-tables -g -fPIC -I/usr/include/gsm' --enable-gpl --disable-x11grab --enable-version3 --enable-pthreads --enable-avfilter --enable-libpulse --enable-libvpx --enable-libopus --enable-libass --disable-libx265 --enable-libmp3lame --enable-libvorbis --enable-libtheora --enable-libspeex --enable-libxvid --enable-libx264 --enable-libschroedinger --enable-libgsm --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-postproc --disable-libdc1394 --enable-librtmp --enable-libfreetype --enable-avresample --enable-libtwolame --enable-libvo-aacenc --enable-gnutls --enable-nonfree --enable-libfdk-aac --enable-libfaac --enable-libopenjpeg --enable-gray --enable-libwebp
  libavutil      52. 92.101 / 52. 92.101
  libavcodec     55. 70.100 / 55. 70.100
  libavformat    55. 49.100 / 55. 49.100
  libavdevice    55. 13.102 / 55. 13.102
  libavfilter     4. 11.102 /  4. 11.102
  libavresample   1.  3.  0 /  1.  3.  0
  libswscale      2.  6.100 /  2.  6.100
  libswresample   0. 19.100 /  0. 19.100
  libpostproc    52.  3.100 / 52.  3.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/tmp/video.mp4':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: isommp42
    creation_time   : 2015-06-24 07:54:29
  Duration: 00:01:04.04, start: 0.000000, bitrate: 1231 kb/s
    Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1280x720 [SAR 1:1 DAR 16:9], 1036 kb/s, 23.98 fps, 23.98 tbr, 24k tbn, 47.95 tbc (default)
    Metadata:
      handler_name    : VideoHandler
    Stream #0:1(und): Audio: aac (mp4a / 0x6134706D), 44100 Hz, stereo, fltp, 192 kb/s (default)
    Metadata:
      creation_time   : 2015-06-24 07:54:29
      handler_name    : IsoMedia File Produced by Google, 5-11-2011
[swscaler @ 0x34bd4a0] deprecated pixel format used, make sure you did set range correctly
Output #0, image2, to '/tmp/images/image-%05d.jpg':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: isommp42
    encoder         : Lavf55.49.100
    Stream #0:0(und): Video: mjpeg, yuvj420p, 1280x720 [SAR 1:1 DAR 16:9], q=2-31, 200 kb/s, 1 fps, 1 tbn, 1 tbc (default)
    Metadata:
      handler_name    : VideoHandler
      encoder         : Lavc55.70.100 mjpeg
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> mjpeg (native))
Press [q] to stop, [?] for help
frame=   67 fps=6.3 q=4.4 Lsize=N/A time=00:01:07.00 bitrate=N/A dup=0 drop=1468    
video:1818kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
d6f897d68eece5d5b721505795b45619f0b3beaac751bce42ce44f39849247fb
7fe0e84fb475e215c8fdd34eaaa8e80d309c071d17bb314740157eb29261b5ba
libdc1394 error: Failed to initialize libdc1394
WARNING: Logging before InitGoogleLogging() is written to STDERR

followed by a ton of this stuff:

}
layer {
  name: "inception_5a/relu_5x5_reduce"
  typ
I0708 01:25:35.832787    22 net.cpp:370] Input 0 -> data
I0708 01:25:35.833078    22 layer_factory.hpp:74] Creating layer conv1/7x7_s2
I0708 01:25:35.833125    22 net.cpp:90] Creating Layer conv1/7x7_s2
I0708 01:25:35.833139    22 net.cpp:410] conv1/7x7_s2 <- data
I0708 01:25:35.833160    22 net.cpp:368] conv1/7x7_s2 -> conv1/7x7_s2
I0708 01:25:35.833247    22 net.cpp:120] Setting up conv1/7x7_s2
I0708 01:25:35.837800    22 net.cpp:127] Top shape: 10 64 112 112 (8028160)
I0708 01:25:35.837831    22 layer_factory.hpp:74] Creating layer conv1/relu_7x7
I0708 01:25:35.837862    22 net.cpp:90] Creating Layer conv1/relu_7x7
I0708 01:25:35.837877    22 net.cpp:410] conv1/relu_7x7 <- conv1/7x7_s2
I0708 01:25:35.837889    22 net.cpp:357] conv1/relu_7x7 -> conv1/7x7_s2 (in-place)
I0708 01:25:35.837904    22 net.cpp:120] Setting up conv1/relu_7x7
I0708 01:25:35.837927    22 net.cpp:127] Top shape: 10 64 112 112 (8028160)
I0708 01:25:35.837939    22 layer_factory.hpp:74] Creating layer pool1/3x3_s2
I0708 01:25:35.837955    22 net.cpp:90] Creating Layer pool1/3x3_s2
I0708 01:25:35.837965    22 net.cpp:410] pool1/3x3_s2 <- conv1/7x7_s2
I0708 01:25:35.837978    22 net.cpp:368] pool1/3x3_s2 -> pool1/3x3_s2

Logs look like

Deepdream ./image-00016.jpg 0 0 inception_4c/output (103, 182, 3) �[2K 0 1 inception_4c/output (103, 182, 3) �[2K 0 2 inception_4c/output (103, 182, 3) �[2K 0 3 inception_4c/output (103, 182, 3) �[2K 0 4 inception_4c/output (103, 182, 3) �[2K 0 5 inception_4c/output (103, 182, 3) �[2K 0 6 inception_4c/output (103, 182, 3) �[2K 0 7 inception_4c/output (103, 182, 3) �[2K 0 8 inception_4c/output (103, 182, 3) �[2K 0 9 inception_4c/output (103, 182, 3) �[2K 1 0 inception_4c/output (144, 255, 3) �[2K 1 1 inception_4c/output (144, 255, 3) �[2K 1 2 inception_4c/output (144, 255, 3) �[2K 1 3 inception_4c/output (144, 255, 3) �[2K 1 4 inception_4c/output (144, 255, 3) �[2K 1 5 inception_4c/output (144, 255, 3) �[2K 1 6 inception_4c/output (144, 255, 3) �[2K 1 7 inception_4c/output (144, 255, 3) �[2K 1 8 inception_4c/output (144, 255, 3) �[2K 1 9 inception_4c/output (144, 255, 3) �[2K 2 0 inception_4c/output (201, 357, 3) �[2K 2 1 inception_4c/output (201, 357, 3) �[2K 2 2 inception_4c/output (201, 357, 3) �[2K 2 3 inception_4c/output (201, 357, 3) �[2K 2 4 inception_4c/output (201, 357, 3) �[2K 2 5 inception_4c/output (201, 357, 3) �[2K 2 6 inception_4c/output (201, 357, 3) �[2K 2 7 inception_4c/output (201, 357, 3) �[2K 2 8 inception_4c/output (201, 357, 3) �[2K 2 9 inception_4c/output (201, 357, 3) �[2K 3 0 inception_4c/output (281, 500, 3) �[2K 3 1 inception_4c/output (281, 500, 3) �[2K 3 2 inception_4c/output (281, 500, 3) �[2K 3 3 inception_4c/output (281, 500, 3) �[2K 3 4 inception_4c/output (281, 500, 3) �[2K 3 5 inception_4c/output (281, 500, 3) �[2K 3 6 inception_4c/output (281, 500, 3) �[2K 3 7 inception_4c/output (281, 500, 3) �[2K 3 8 inception_4c/output (281, 500, 3) �[2K 3 9 inception_4c/output (281, 500, 3) �[2K Error Code is 0 Just created outputs/./image-00016.jpg Deepdream ./image-00061.jpg

lastly, after canceling it

KeyboardInterrupt
^Cffmpeg version git-2014-07-28-a776238 Copyright (c) 2000-2014 the FFmpeg developers
  built on Jul 28 2014 03:08:49 with gcc 4.8 (Ubuntu 4.8.2-19ubuntu1)
  configuration: --shlibdir=/usr/lib64 --prefix=/usr --mandir=/usr/share/man --libdir=/usr/lib64 --enable-static --extra-cflags='-fmessage-length=0 -grecord-gcc-switches -fstack-protector -O2 -Wall -D_FORTIFY_SOURCE=2 -funwind-tables -fasynchronous-unwind-tables -g -fPIC -I/usr/include/gsm' --enable-gpl --disable-x11grab --enable-version3 --enable-pthreads --enable-avfilter --enable-libpulse --enable-libvpx --enable-libopus --enable-libass --disable-libx265 --enable-libmp3lame --enable-libvorbis --enable-libtheora --enable-libspeex --enable-libxvid --enable-libx264 --enable-libschroedinger --enable-libgsm --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-postproc --disable-libdc1394 --enable-librtmp --enable-libfreetype --enable-avresample --enable-libtwolame --enable-libvo-aacenc --enable-gnutls --enable-nonfree --enable-libfdk-aac --enable-libfaac --enable-libopenjpeg --enable-gray --enable-libwebp
  libavutil      52. 92.101 / 52. 92.101
  libavcodec     55. 70.100 / 55. 70.100
  libavformat    55. 49.100 / 55. 49.100
  libavdevice    55. 13.102 / 55. 13.102
  libavfilter     4. 11.102 /  4. 11.102
  libavresample   1.  3.  0 /  1.  3.  0
  libswscale      2.  6.100 /  2.  6.100
  libswresample   0. 19.100 /  0. 19.100
  libpostproc    52.  3.100 / 52.  3.100
[image2 @ 0x33380c0] Could find no file with path '/tmp/outputs/image-%05d.jpg' and index in the range 0-4
/tmp/outputs/image-%05d.jpg: No such file or directory
cp: cannot stat ‘/tmp/out.mp4’: No such file or directory

Add code to generate zoom+deepdream animated gif

Simple command line script for doing this with imagemagick would come in handy for those that haven't done it before.

Add link to gallery of example outputs

I used your excellent docker container to create an example image for each 'end' layer selection.

http://imgur.com/a/yMCGZ

Would you consider including that link into your README so others can see how each layer option ultimately affects the output?

How to Continue Iterations

Hello,

I can't seem to figure out how to continue the iterations to create the zooming in loop, even after un-comenting out the frame counter and relevant code at the end of the deepdream.py script

Would love to be able to write in a frame count in the json settings file, is this possible?

Only works on some input images

I can only get the program to run on certain images.

Does not work (error code 1):
filename: noise.jpg

Works but produces output with incorrect file extension:
filename: test.png (output file is a jpg with .png extension)

Consider adding instructions on how to increase boot2docker memory limit

I'm familiar with docker, but not boot2docker on OSX, and it took me quite a while to figure out how to increase the VM's memory size. You reference that it might be necessary to up the RAM for processing of larger images, so it would be helpful to have a quick link on how to change the configuration to match.

I followed the instructions here: https://docs.docker.com/articles/b2d_volume_resize/

Specifically,

1. Initialise a default file to customize
   boot2docker config > ~/.boot2docker/profile command.
2. Edit the file and change from 2 Gig to 6 Gig
   Memory = 2048
   to
   Memory = 6144
3. Run the following commands:

$ boot2docker poweroff
$ boot2docker destroy
$ boot2docker init
$ boot2docker up

Is it possible enable multithreading in deepdream function?

I've noticed that the actual deepdream function that does the work actually runs on a single thread. So no matter how powerful multicore processor you got, it will only use a very small portion of it.
I am a newbie in this field, can someone tell if its possible to have a multithreaded function ?

I executed htop inside the compute container and i got this... notice only core 1 is used and others are idle.

  1  [|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||100.0%]     5  [                                                                     0.0%]
  2  [                                                                     0.0%]     6  [                                                                     0.0%]
  3  [                                                                     0.0%]     7  [                                                                     0.0%]
  4  [                                                                     0.0%]     8  [                                                                     0.0%]
  Mem[|||||||||||||||||||||||||||||||||||||||||||||||||             1140/2001MB]     Tasks: 7, 1 thr; 2 running
  Swp[|||||||||||||||||                                               103/460MB]     Load average: 1.17 1.06 0.87 
                                                                                     Uptime: 05:00:11

  PID USER      PRI  NI  VIRT   RES   SHR S CPU% MEM%   TIME+  Command
  528 root       20   0 1633M 1037M 42900 R 100. 51.8  0:42.06 python deepdream.py
  544 root       20   0 23020  3292  2632 R  0.7  0.2  0:00.01 htop
    1 root       20   0 17960     0     0 S  0.0  0.0  0:00.07 /bin/bash -c cd /opt/deepdream && ./process_images.sh 2>&1 > log.html
    6 root       20   0 17964     0     0 S  0.0  0.0  0:00.00 /bin/bash ./process_images.sh
    7 root       20   0 17980     0     0 S  0.0  0.0  0:00.00 /bin/bash ./process_images_once.sh
    9 root       20   0 17984  1916  1784 S  0.0  0.1  0:00.04 /bin/bash ./process_images_once.sh
  529 root       20   0 1633M 1037M 42900 S  0.0 51.8  0:00.00 python deepdream.py
  530 root       20   0 18168  3300  2844 S  0.0  0.2  0:00.09 bash





F1Help  F2Setup F3SearchF4FilterF5Tree  F6SortByF7Nice -F8Nice +F9Kill  F10Quit

Followed digitalocean instructions, getting No such file or directory

1124 07:35:09.549391 72 upgrade_proto.cpp:618] Attempting to upgrade input file specified using deprecated V1LayerParameter: ../caffe/models/bvlc_googlenet/bvlc_googlenet.caffemodel
I1124 07:35:09.681547 72 upgrade_proto.cpp:626] Successfully upgraded file specified using deprecated V1LayerParameter
/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)
./process_images_once.sh: line 24: 72 Killed python deepdream.py
cp: cannot stat 'output.jpg': No such file or directory
rm: cannot remove 'output.jpg': No such file or directory
ffmpeg version git-2014-07-28-a776238 Copyright (c) 2000-2014 the FFmpeg developers
built on Jul 28 2014 03:08:49 with gcc 4.8 (Ubuntu 4.8.2-19ubuntu1)
configuration: --shlibdir=/usr/lib64 --prefix=/usr --mandir=/usr/share/man --libdir=/usr/lib64 --enable-static --extra-cflags='-fmessage-length=0 -grecord-gcc-switches -fstack-protector -O2 -Wall -D_FORTIFY_SOURCE=2 -funwind-tables -fasynchronous-unwind-tables -g -fPIC -I/usr/include/gsm' --enable-gpl --disable-x11grab --enable-version3 --enable-pthreads --enable-avfilter --enable-libpulse --enable-libvpx --enable-libopus --enable-libass --disable-libx265 --enable-libmp3lame --enable-libvorbis --enable-libtheora --enable-libspeex --enable-libxvid --enable-libx264 --enable-libschroedinger --enable-libgsm --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-postproc --disable-libdc1394 --enable-librtmp --enable-libfreetype --enable-avresample --enable-libtwolame --enable-libvo-aacenc --enable-gnutls --enable-nonfree --enable-libfdk-aac --enable-libfaac --enable-libopenjpeg --enable-gray --enable-libwebp
libavutil 52. 92.101 / 52. 92.101
libavcodec 55. 70.100 / 55. 70.100
libavformat 55. 49.100 / 55. 49.100
libavdevice 55. 13.102 / 55. 13.102
libavfilter 4. 11.102 / 4. 11.102
libavresample 1. 3. 0 / 1. 3. 0
libswscale 2. 6.100 / 2. 6.100
libswresample 0. 19.100 / 0. 19.100
libpostproc 52. 3.100 / 52. 3.100
[image2 @ 0x35c00c0] Could find no file with path '/tmp/outputs/image-%05d.jpg' and index in the range 0-4
/tmp/outputs/image-%05d.jpg: No such file or directory
cp: cannot stat '/tmp/out.mp4': No such file or directory

get it running on Windows

Could anyone point me in the right direction? I don't seem to find the right approach to get it fully running.

Putting Different Images in input folder and still getting same result

I add a group of different jpgs into the inputs folder. The script works and generates dream imagery but while the filenames match the image the name, the image generated is always the same (usually matching the first image I upload)

Images aren't automatically being processed

I followed your directions for Digital Ocean, Ubuntu with Docker installed. I scp'ed the image to my server but noticed that nothing was displaying.. so I manually ran process_images.sh and got this error:

./process_images.sh: 6: cd: can't cd to /opt/deepdream
./process_images.sh: 10: cd: can't cd to /opt/deepdream/inputs
/root/clouddream/deepdream/inputs
File is input.jpg
Deepdream input.jpg
cp: cannot stat ‘inputs/input.jpg’: No such file or directory
pwd is /root/clouddream/deepdream/inputs
python: can't open file 'deepdream_single.py': [Errno 2] No such file or directory
cp: cannot stat ‘output.jpg’: No such file or directory
./process_images.sh: 10: cd: can't cd to /opt/deepdream/inputs
/root/clouddream/deepdream
File is 6031848_1547091962_98large.jpg
Deepdream 6031848_1547091962_98large.jpg
pwd is /root/clouddream/deepdream
Traceback (most recent call last):
  File "deepdream_single.py", line 3, in <module>
    import numpy as np
ImportError: No module named numpy
cp: cannot stat ‘output.jpg’: No such file or directory

Is there something I'm doing wrong here? Sorry, I'm new with Linux, Python and such.

AWS IP Does not show images or video.

When I type in http://1.2.3.4:8000, for example, I never see the frames that are being processed. I cannot see the images as well. Does anyone know why? (I am not typing in 1.2.3.4, obviously :))

Add a way to generate 30fps videos.

Would be pretty cool for short videos.

Docs for using clouddream locally with boot2docker?

Any chance we could get docs on running this locally in our own instances of docker?

Pretty please?

:)

Thanks!!

move deepdream.vision.ai specific stuff into separate branch

There should be a separate branch which contains customizations used in deepdream.vision.ai, which I doubt anybody here actually wants:

1. Google Analytics embedded in the homepage
2. The "Fork me on Github" button

root password on scp image upload

seems to want a root password on scp image upload... is this a permission issue?