Installation and Setup

Python compatibility: 3.4-3.6 | 64bit

Installation

via pip

pip install imagepypelines --user

dependencies

when running natively, ImagePypelines requires opencv and tensorflow to be installed on your machine

tensorflow

pip install tensorflow --user
# OR if you have a GPU
# pip install tensorflow-gpu --user

opencv

we strongly recommend that you build opencv from source. However unofficial bindings for Opencv can be installed with

pip install opencv-python --user

(while we haven’t encountered many problems with these unofficial bindings, they will likely not be as optimized and we do not guarantee support)

The ImagePypelines shell

ImagePypelines provides docker images (virtual environments) which contain all the libraries and tools you’ll need to develop and run ImagePypeline’s code

  1. Simply install Docker
  2. type imagepypelines shell into your terminal!

You be will launched into a virtual environment that is guaranteed to work hassle-free without installing your dependencies.

ImagePypelines Shell