Change seed

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

         network_settings=new_networksettings,
         max_steps=3000,
     )
-    check_environment_trains(env, {BRAIN_NAME: config})
+    check_environment_trains(env, {BRAIN_NAME: config}, training_seed=1212)
 
 
 @pytest.mark.parametrize("action_sizes", [(0, 1), (1, 0)])