This chapter will help you to configure the backend implementation of Keras. It includes the installation of TensorFlow and Theano. Let us install these two libraries and configure them for the backend implementation:
To install TensorFlow, you will be required to use the following command:
pip install TensorFlow
After the successful installation of TensorFlow, you are required to do the following configurations in the keras.json file:
{
"image_data_format": "channels_last",
"epsilon": 1e-07, "floatx": "float32", "backend": "tensorflow"
}
After that, you can install Theano with the below command:
pip install theano
Next, the keras.json file will have the following configurations:
{
{
"image_data_format": "channels_last",
"epsilon": 1e-07,
"floatx": "float32",
"backend": "theano"
}
Now, you’re all done with installation and configurations. And you can start Keras with the help of the command given below:
>>>import keras a krs