July 2018
Intermediate to advanced
254 pages
8h 12m
English
It feels great, doesn't it? You have now developed an agent that can learn to play any Atari game and get better at it by itself! Once you have your agent trained on any Atari game, you can use the test mode of the script to test the agent's performance based on its learning so far. You can enable the test mode by using the --test argument in the deep_q_learner.py script. It is useful to enable rendering of the environment, too, so that you can visually see (apart from the rewards printed on the console) how the agent is performing. As an example, you can test the agent in the Seaquest Atari game using the following command:
(rl_gym_book) praveen@ubuntu:~/HOIAWOG/ch6$ python deep_Q_learner.py ...Read now
Unlock full access