You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The implementation does not have a way of assigning the max_steps when creating an MPE environment. I believe this is a really small change, but it needs to be made for every simple_*.py file.
In addition, the current implementation of the baseline IPPO for MPE use higher max_steps which makes the environment start over in a single update and there is no reset between updates. I suspect this is the case for other baseline algorithms.
The text was updated successfully, but these errors were encountered:
Hi, the argument that you mention is for the base class of the MPE environments and it is not accessible when creating one of the MPE environments.
For example, the class SimpleReferenceMPE does not have this argument and is set by default to 25 by the base class. I can change it afterwards calling something like env.max_steps=new_value but this is cumbersome and I feel that it should be possible to set it when creating the environment.
The implementation does not have a way of assigning the max_steps when creating an MPE environment. I believe this is a really small change, but it needs to be made for every simple_*.py file.
In addition, the current implementation of the baseline IPPO for MPE use higher max_steps which makes the environment start over in a single update and there is no reset between updates. I suspect this is the case for other baseline algorithms.
The text was updated successfully, but these errors were encountered: