looks like target-network and online-network gets updated at the same frequency UPDATE_FREQ

Open anirbanchowdhury9862 opened this issue 5 years ago • 3 comments

May 15 '20 02:05 anirbanchowdhury9862

Thank you for spotting this! I've just made the fix in a local file, and am currently testing it, with the target-network being updated every 10,000 steps. I'll get back to you when it's done.

May 15 '20 12:05 sebtheiler

I apologize that this has taken me so long, the past few months have been hectic. I've created a new branch dqn-fix with the changes. I haven't had time to thoroughly test the results and make sure the DQN obtains a similar score, but the code is there, and from my testing so far, it's working as expected. I'll be away from my desktop for a month or so and won't be able to do any more testing for the moment, which is why I'm uploading the files now. I'll leave this issue open until I can test the code to completion.

Jun 27 '20 00:06 sebtheiler

Great.

Jul 03 '20 03:07 anirbanchowdhury9862