Code for a customized version of the Monster Kong environment, which was originally provided by https://github.com/ntasfi/PyGame-Learning-Environment.

We consider 20 custom maps for testing multi-task and continual learning settings when we move the goal location and location of coins that the agent is rewarded for collecting. The first 10 maps focus on jumping between platforms while the second 10 maps focus on climbing ladders.