Increasing steps on asymmetric ghost test (#3802)

ml-agents/mlagents/trainers/tests/test_simple_rl.py

         [BRAIN_NAME + "?team=0", brain_name_opp + "?team=1"], use_discrete=use_discrete
     )
     override_vals = {
-        "max_steps": 2000,
+        "max_steps": 4000,
-            "save_steps": 5000,
-            "swap_steps": 5000,
-            "team_change": 2000,
+            "save_steps": 10000,
+            "swap_steps": 10000,
+            "team_change": 4000,
         },
     }
     config = generate_config(PPO_CONFIG, override_vals)