-
Notifications
You must be signed in to change notification settings - Fork 41
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
Adding multidiscrete feature #328
base: master
Are you sure you want to change the base?
Conversation
Hi, @Jogima-cyber ! Long time no see, and thanks for your great suggestion! |
I'm currently working on it ! First thing I'm doing right now is integrating a home made simple multidiscrete environment into HandyRL in order to test my upcoming multidicrete integration. I'll push it as soon as I'm finished (it's gonna include a non-multidiscrete option). |
@Jogima-cyber In this branch, each agent can output |
Thank you @YuriCat, that's almost exactly what I needed. I'm working on your fork, I just need to add a moment "unit_mask", because some of my independent action sets may be empty depending on the state (meaning no action can be taken among the actions of the action set). |
Indeed! We need |
Btw. I've output the shapes of some vectors and in line feature/multi_unit/train.py:l204 I've got shapes that I don't think are normal :
|
@Jogima-cyber |
Yeah I found where to add a |
@Jogima-cyber Updated my branch feature/numti_unit. |
Hello there ! I'd like to add the capability to handle multidiscrete action space. Here is how I would do that :