I want to use two different reinforcement algorithms for each team in my basic game. In the same gaming time, one team will run with the PPO and the other one with the SAC. Algorithms are built-in in the unity toolkit. Is it possible?
You can configure it to train multiple RL models at the same time. The StrikerVsGoalie example does this. It uses two POCA trainers but I'm not aware of any reason they couldn't be different.