HumanCompatibleAI/overcooked_ai

How to repeat the Near-optimal model-based planner and Planning experiments(in old dynamics))？

wittycheng opened this issue · 4 comments

wittycheng commented

I am wondering how to use the deprecated MediumLevelPlanner to repeat. I met this problem
image
image

wittycheng commented

and many unfound functions in class Heuristic

jyan1999 commented

Hi,

If you have specific needs for this function, I'd recommend going back and using an older commit before this method was deprecated and following the instructions included in that commit. For example the DELIVERY_REW_THRES can be found in this commit. Note we are unlikely to be able to offer support for deprecated methods and classes.

  • 👍1
wittycheng commented

thank you， I will take a try

micahcarroll commented

Just to add on to that, if you want to reproduce the old experiments exactly as in the paper I'd use this tagged version

  • 👍1