Dodgeball config update

config/ppo/DodgeBall.yaml

   DodgeBall:
     trainer_type: ppo
     hyperparameters:
-      batch_size: 128
-      buffer_size: 2048
+      batch_size: 1024
+      buffer_size: 10240
       learning_rate: 0.0003
       beta: 0.01
       epsilon: 0.2
 #            encoding_size: 256
 #            learning_rate: 0.0003
     keep_checkpoints: 5
-    max_steps: 2000000
+    max_steps: 20000000
     time_horizon: 64
     summary_freq: 5000
     threaded: true