[change] Change warning language when adversarial scene is used without self-play (#3561)

ml-agents/mlagents/trainers/ppo/trainer.py

         """
         if self.policy:
             logger.warning(
-                "add_policy has been called twice. {} is not a multi-agent trainer".format(
+                "Your environment contains multiple teams, but {} doesn't support adversarial games. Enable self-play to \
+                    train adversarial games.".format(
                     self.__class__.__name__
                 )
             )

ml-agents/mlagents/trainers/sac/trainer.py

         """
         if self.policy:
             logger.warning(
-                "add_policy has been called twice. {} is not a multi-agent trainer".format(
+                "Your environment contains multiple teams, but {} doesn't support adversarial games. Enable self-play to \
+                    train adversarial games.".format(
                     self.__class__.__name__
                 )
             )