Reinforcement learning (open_ai) Evaluation in real robot

Hi all. do you know how to test(evaluate) the trained agent in DeepQN turtlebot on real robot or gazebo?
i don’t know how to load the saved model and tested in real robot or simulation


In the course “Use openAI with ROS” there are examples on how to load and test" a deepQ network and Her created previously ( chapter 6 )

if you used Tensorflow for your model you can go to the Tensorflow page and search for load models. it is just one o perhaps 2 lines of code