-
Notifications
You must be signed in to change notification settings - Fork 1.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Acer #231
base: master
Are you sure you want to change the base?
Acer #231
Conversation
Add memory and check forward. Compile successful
Make episodic memory efficient
@matthiasplappert Current implementation supports single agent learning for ACER. Do you think we should add synchronous agent learning approach as implemented by OpenAI gym? @RaphaelMeudec Let me know if your views as well. |
Note don't share parameters in network
Do not share network parameters
I guess we need more reviewers. I can volunteer to review and test. |
@AloshkaD the current implementation supports simple toy examples. It will really be helpful if you volunteer to review this PR. You could also work to add support for Atari games in ACER. Let me know if you need any explanation in the code. Thanks |
This is implementation of ACER algorithm using keras-rl.
Paper : https://arxiv.org/pdf/1611.01224.pdf