userGuide

[reillydonovan]

Hello!
I was wondering if you could include a more detailed user guide.
I have successfully installed the deepdream notebook and can get it to apply different treatments to my own photos.
Now for deepDreamVideo...

r2d2@r2d2-VirtualBox:~/Downloads/DeepDreamVideo$ ./1_movie2frames [ffmpeg|avconv] BUBBLES.mov movFrames
When I enter the following command for the 25fps conversion, terminal returns the following:
bash: ./1_movie2frames: No such file or directory
No command 'avconv]' found, did you mean:
Command 'avconv' from package 'libav-tools' (universe)
avconv]: command not found

If I enter /.1_movie2frames.sh it returns:
please provide the moviename and directory where to store the frames
./1_movie2frames [ffmpeg|avconv] [movie.mp4] [directory]

I do have lib-avconv installed.

Also, for the following command this is what is returned:
python 2_dreaming_time.py -i frames -o processed --gpu 0
r2d2@r2d2-VirtualBox:~/Downloads/DeepDreamVideo$ python 2_dreaming_time.py -i frames -o processed --gpu 0
usage: 2_dreaming_time.py [-h] -i INPUT -o OUTPUT [-d] [-g] [-t MODEL_PATH]
[-m MODEL_NAME] [-p PREVIEW] [-oct OCTAVES]
[-octs OCTAVESCALE] [-itr ITERATIONS] [-j JITTER]
[-z ZOOM] [-s STEPSIZE] [-b BLEND]
[-l LAYERS [LAYERS ...]]
2_dreaming_time.py: error: unrecognized arguments: 0

does this have to be executed in the frames folder? Does it produce a new frame with a processed extension or a new folder?

Any help or guidance on how to properly execute these commands would be very appreciated!
Thank you so much!

[graphific]

hey first of all try:
./1_movie2frames avconv BUBBLES.mov movFrames
you're supposed to provide ffmpeg OR avconv as an argument.

any luck getting further with that?

[reillydonovan]

I converted the video to frames via blender for the time being to move on to step two.
I dropped those stills into the frames folder and executed the python script and it returned cannot launch GPU with CPU only so I'm remaking caffe without CPU only... perhaps there is a way to run your code under CPU only?

I tried your recommendation and this is what it returned:

r2d2@r2d2-VirtualBox:/Downloads/DeepDreamVideo$ ./1_movie2frames avconv BUBBLES.mov movFrames
bash: ./1_movie2frames: No such file or directory
r2d2@r2d2-VirtualBox:/Downloads/DeepDreamVideo$ ls
1_movie2frames.sh 3_frames2movie.sh frames2gif.sh movFrames README.md
2_dreaming_time.py frames LICENSE processed tmp.prototxt

[graphific]

you forgot the extension: what about
./1_movie2frames.sh avconv BUBBLES.mov movFrames

[reillydonovan]

r2d2@r2d2-VirtualBox:~/Downloads/DeepDreamVideo$ ./1_movie2frames.sh avconv BUBBLES.mov movFrames
./1_movie2frames.sh: line 12: ffmpeg: command not found

I guess Ubuntu 14.04 doesn't have ffmpeg... but shouldn't avconv cover for that?
Thanks for your help!

[reillydonovan]

is there a way to run the 2_dreaming_time.py script as CPU only?

[rosshamish]

Seems like it'll run CPU only unless you pass the -g flag (from #14)

Edit: and from the readme:

(don't forget the --gpu flag if you got a gpu to run on, where 0 is the index of 
the gpu you'd like to use if you have more than 1)

[reillydonovan]

great thanks, that works! Now the next problem...
It starts and then it crashes... before it returned something about not allocating enough memory and doing a core dump...

Processing frame #1
/home/r2d2/anaconda/lib/python2.7/site-packages/scipy/ndimage/interpolation.py:549: 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)
(0, 0, 'inception_4d/output', (320, 569, 3))
(0, 1, 'inception_4d/output', (320, 569, 3))
(0, 2, 'inception_4d/output', (320, 569, 3))
(0, 3, 'inception_4d/output', (320, 569, 3))
(0, 4, 'inception_4d/output', (320, 569, 3))
(1, 0, 'inception_4d/output', (480, 853, 3))
(1, 1, 'inception_4d/output', (480, 853, 3))
(1, 2, 'inception_4d/output', (480, 853, 3))
(1, 3, 'inception_4d/output', (480, 853, 3))
(1, 4, 'inception_4d/output', (480, 853, 3))
(2, 0, 'inception_4d/output', (720, 1280, 3))
Killed

I'm running a virtual machine so I think I might not have enough available ram or storage for the script to complete or since I'm doing CPU only its running out of buffer space. Getting closer!

[trotskylenin]

I'm having the exactly same problem as @reillydonovan

[trotskylenin]

I'm using an Intel i3 core processor to do the job, no GPU card on my notebook with Ubuntu 14.04. I think the problem has to do with CPU lacking of 'horsepower' 👎 or something like that.

[trotskylenin]

I opened another issue with this @reillydonovan --> #20

[graphific]

As I understand, its possible to run the code now with just cpu, correct?

as for the task being killed after the first frame as @reillydonovan writes, let's continue at #20.

[graphific]

@trotskylenin @reillydonovan Ive pulled a new PR for the cpu behavior. its not needed to give an argument to the -g flag anymore.

could you test if the code works now with cpu only?

[graphific]

Ive started a wiki with advise as well, feel free to add to it: https://github.com/graphific/DeepDreamVideo/wiki

[Acanterelle]

Can 2_dreaming_time.py run with a boot2docker dreamer such as ryankennedy's? I've managed to get 1_movies2frames.sh working, but the output seems a little short. For a 6 second video I only get 50 frames.

Thanks

[rosshamish]

@Acanterelle I'm going to try to dockerize this tonight