Fixing RL test.

Author: Błażej O

tensor2tensor/models/research/rl.py

@@ -52,7 +52,7 @@ def ppo_base_v1():
   hparams.add_hparam("save_models_every_epochs", 30)
   hparams.add_hparam("optimization_batch_size", 50)
   hparams.add_hparam("max_gradients_norm", 0.5)
-  hparams.add_hparam("simulated_environment", True)
+  hparams.add_hparam("simulated_environment", False)
   return hparams