changed learning rate tuning

reinforce_tune_lr.py

@@ -2,8 +2,8 @@ import os
 from itertools import product
 
 parameters = dict(
-    policy_lr = [0.1, 0.01],
-    value_lr = [0.1, 0.01]
+    policy_lr = [0.01, 0.001, 0.0001, 0.00001],
+    value_lr = [0.01, 0.001, 0.0001, 0.00001]
 )
 
 param_values = [v for v in parameters.values()]