Benson V7 be training

MLAgents/results/BensonV7/configuration.yaml

@@ -42,11 +42,11 @@ behaviors:
     summary_freq: 10000
     threaded: false
     self_play:
-      save_steps: 20000
-      team_change: 100000
-      swap_steps: 10000
+      save_steps: 40000
+      team_change: 200000
+      swap_steps: 20000
       window: 10
-      play_against_latest_model_ratio: 0.5
+      play_against_latest_model_ratio: 0.666
       initial_elo: 1200.0
     behavioral_cloning: null
 env_settings:
@@ -72,8 +72,8 @@ checkpoint_settings:
   run_id: BensonV7
   initialize_from: null
   load_model: false
-  resume: false
-  force: true
+  resume: true
+  force: false
   train_model: false
   inference: false
   results_dir: results