Keras is a very popular deep learning framework on its own and it is heavily used by newcomers looking to learn about the basics of constructing networks. The framework is considered very high-level and abstracts most of the inner details of constructing networks. It only goes to assume that an RL framework built with Keras would attempt to do the same thing.
This example is dependent on the version of Keras and TensorFlow and may not work correctly unless the two can work together. If you encounter trouble, try installing a different version of TensorFlow and try again.
To run this example, we will start by doing the installation and all of the setup in this exercise:
- To install Keras, you should create a new virtual environment using Python 3.6 and use pip to install it along with the keras-rl framework. The commands to do all of this on Anaconda are...
