tennis config

config/trainer_config.yaml

     hidden_units: 256
     time_horizon: 1000
     self_play:
-        window: 10
-        play_against_current_self_ratio: 0.5
+        window: 100
+        play_against_current_self_ratio: 0.2
         save_steps: 50000
         swap_steps: 50000