Lately, I have implemented DQN for Atari Breakout. Here is the code:


I have trained the agent for over 1500 episodes, but the training leveled off at around 5 as score. Could someone look at the code and point out a few things that should be corrected?

enter image description here

Actually, the training is not going more than 5 in average score. Is there a way to improve my performance?

  • $\begingroup$ Have you look at the original DQN paper? How long have they trained their network for? $\endgroup$
    – nbro
    Apr 10, 2020 at 14:41
  • $\begingroup$ Yeah, I have seen the original paper. I trained it for 3 to 5 hours $\endgroup$
    – jgauth
    Apr 10, 2020 at 14:42
  • $\begingroup$ And how long have the original authors of DQN trained their network for? Probably a lot more than 3-5 hours. And in terms of episodes? $\endgroup$
    – nbro
    Apr 10, 2020 at 14:43
  • $\begingroup$ ''One epoch corresponds to 50000 minibatch weight updates or roughly 30 minutes of training time''. They achieved average reward of 1800, and in my case I get at most 5. Where am I wrong? Will I achieve 1600 or 1700 in average reward if I wait long enough. Right now, the code have some difficulties to just break through 5 in average reward $\endgroup$
    – jgauth
    Apr 10, 2020 at 14:55
  • 1
    $\begingroup$ It's been a long time since I read that paper, but I was planning to read it again in the next days. Maybe I will have a look at it later. $\endgroup$
    – nbro
    Apr 10, 2020 at 15:36


You must log in to answer this question.

Browse other questions tagged .