-
Notifications
You must be signed in to change notification settings - Fork 271
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
Disable the done signal. #45
Conversation
Actually, this change is still largely incomplete, since I need to update all of the environment dictionaries. I'll probably have time to do that in a week's time. |
944b348
to
f89dbb4
Compare
I believe there was some discussion that came to the conclusion that adding this option is fine? If so, I believe this change can be merged in its current state. |
I think this could be made simpler by simply disabling it universally. We should also raise an exception in each env if someone steps past 150. |
Perhaps? I can certainly imagine someone wanting to be able to easily preserve the old behavior.
I added that now. Since we probably want that check to be done in a base class method, I don't think there's much more simplicity to be had. |
I guess the missing context is that I discussed this with the authors last week and we decided that the simplest thing to do is to never send a done signal. This is consistent with the CoRL publication, and the current version is inconsistent. Those needing a time limit implemented via termination (e.g. some on-policy implementations) will have to add it themselves. I apologize for not specifically communicating about this (one of those many TODOs which never got visited). |
0b28e1e
to
0f40380
Compare
477ce48
to
e47607a
Compare
Also, throw an exception if the done signal is ignored.
e47607a
to
ec4f9ef
Compare
@krzentner is this ready to merge? |
Yes. It leaves the path length alone (150 in some envs, 200 in others), removes the done signal, and throws an exception when the path length is exceeded. |
Disable the done signal (Farama-Foundation#45)
This fixes #43 in what I think is the cleanest way possible. There should probably be more discussion before we officially make a v2, but this might be of use to some people immediately.
This is definitely of interest to @avnishn and @lywong92 .