HandyRL is a handy and simple framework based on Python and PyTorch for distributed reinforcement learning that is applicable to your own environments.
observation config should be able to set both by each environment and each agent.
For example, in a Geister environment, it is normal to set up the ability to observe on the opponent's turn, but it is possible for some agents to play without observing.
observationconfig should be able to set both by each environment and each agent.For example, in a Geister environment, it is normal to set up the ability to observe on the opponent's turn, but it is possible for some agents to play without observing.