fix advanced vis encoder simple rl

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

         PPO_TORCH_CONFIG,
         hyperparameters=new_hyperparams,
         network_settings=new_networksettings,
-        max_steps=700,
+        max_steps=900,
         summary_freq=100,
     )
     # The number of steps is pretty small for these encoders