reduce max_steps for Gridworld (#2973)

config/sac_trainer_config.yaml

     init_entcoef: 0.5
     buffer_init_steps: 1000
     buffer_size: 50000
-    max_steps: 5.0e5
+    max_steps: 50000
     summary_freq: 2000
     time_horizon: 5
     reward_signals:

config/trainer_config.yaml

     hidden_units: 256
     beta: 5.0e-3
     buffer_size: 256
-    max_steps: 5.0e5
+    max_steps: 50000
     summary_freq: 2000
     time_horizon: 5
     reward_signals: