<languages />
[[Category:Software]][[Category:AI and Machine Learning]]
"Keras is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano."<ref>https://keras.io/</ref>

If you are porting a Keras program to one of our clusters, you should follow [[Tutoriel Apprentissage machine/en|our tutorial on the subject]].

==Installing==

#Install [[TensorFlow]], CNTK, or Theano in a Python [[Python#Creating_and_using_a_virtual_environment|virtual environment]].
#Activate the Python virtual environment (named <tt>$HOME/tensorflow</tt> in our example).
#:{{Command2|source $HOME/tensorflow/bin/activate}}
#Install Keras in your virtual environment.
#:{{Command2
|prompt=(tensorflow)_[name@server ~]$
|pip install keras}}

=== R package ===

This section details how to install Keras for R and use TensorFlow as the backend.

#Install TensorFlow for R by following [[Tensorflow#R_package | these instructions]].
#Follow the instructions from the parent section.
#Load the required modules. 
#:{{Command2|module load gcc/7.3.0 r/3.5.2}}
# Launch R.
#:{{Command2|R}}
#In R, install the Keras package with <code>devtools</code>. 
#:<syntaxhighlight lang='r'>
devtools::install_github('rstudio/keras')
</syntaxhighlight>


You are then good to go. Do not call <code>install_keras()</code> in R, as Keras and TensorFlow have already been installed in your virtual environment with <code>pip</code>. To use the Keras package installed in your virtual environment, enter the following commands in R after the environment has been activated.
<syntaxhighlight lang='r'>
library(keras)
use_virtualenv(Sys.getenv('VIRTUAL_ENV'))
</syntaxhighlight>

== References ==