This framework is tested to work under Python 3.6.
The required dependencies are NumPy >= 1.10, joblib >= 0.9. You also need keras or you can write your own learning algorithms using your favorite deep learning framework.
For running some of the examples, Matplotlib >= 1.1.1 is required. You also sometimes need to install specific dependencies (e.g. for the atari games, you need to install ALE >= 0.4).
We recommend to use the bleeding-edge version and to install it by following the Developer install instructions. If you want a simpler installation procedure and do not intend to modify yourself the learning algorithms etc., you can look at the User install instructions.
Developer install instructions¶
As a developer, you can set you up with the bleeding-edge version of DeeR with:
git clone -b master https://github.com/VinF/deer.git
Assuming you already have a python environment with
pip, you can automatically install all the dependencies (except specific dependencies that you may need for some examples) with:
pip install -r requirements.txt
And you can install the framework as a package using the mode
develop so that you can make modifications and test without having to re-install the package.
python setup.py develop
User install instructions¶
You can install the framework with pip:
pip install deer
For the bleeding edge version (recommended), you can simply use
pip install git+git://github.com/VINF/deer.git@master