DeepDream Video Generator with ffmpeg, pycaffe
Python Shell
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


I'm hired for people to fulfill their fantasies. Their **DEEP DARK FANTASIES**. - Van Darkholme

Sorry, It's just a joke. It makes DeepDream video with using ffmpeg and caffe. This project is targeted at Ubuntu 14.04 LTS.


Install CUDA 7.5 yourself. Then our Install Script will install every dependencies.

How to Use?

  • In first, Run Installer Script.
$ ./
  • After setup, Set your configure file. If you are using vim,
$ vi configure.conf

or if you are using emacs,

$ sudo apt-get remove emacs
$ vi configure.conf
  • And, Run main script.
$ ./

Configure Options

  • INPUT_FOLDER: Temporary Input Folder for Extracting video.
  • OUTPUT_FOLDER: Temporary Output Folder for Processing frames.
  • VIDEO_FILE: Video file for process.
  • OUTPUT_FILE: Output video file.
  • FRAMERATE: Video's Framerate
  • MODEL_IMAGE: Deep Dream Model Image
  • THREADS: CPU's threads. (Only for ffmpeg)
  • ITERATIONS: Iterations when processing images.
  • OCTAVES: Octaves when processing images.
  • OCTAVES_SCALE: Octaves Scale when processing images.