fixed the LSTM middleware initialization

agents/ddpg_agent.py

@@ -37,6 +37,8 @@ class DDPGAgent(ActorCriticAgent):
         self.q_values = Signal("Q")
         self.signals.append(self.q_values)
 
+        self.reset_game(do_not_reset_env=True)
+
     def learn_from_batch(self, batch):
         current_states, next_states, actions, rewards, game_overs, _ = self.extract_batch(batch)
 

agents/policy_optimization_agent.py

@@ -47,6 +47,8 @@ class PolicyOptimizationAgent(Agent):
         self.entropy = Signal('Entropy')
         self.signals.append(self.entropy)
 
+        self.reset_game(do_not_reset_env=True)
+
     def log_to_screen(self, phase):
         # log to screen
         if self.current_episode > 0:

agents/ppo_agent.py

@@ -45,6 +45,8 @@ class PPOAgent(ActorCriticAgent):
         self.unclipped_grads = Signal('Grads (unclipped)')
         self.signals.append(self.unclipped_grads)
 
+        self.reset_game(do_not_reset_env=True)
+
     def fill_advantages(self, batch):
         current_states, next_states, actions, rewards, game_overs, total_return = self.extract_batch(batch)
 

agents/value_optimization_agent.py

@@ -28,6 +28,8 @@ class ValueOptimizationAgent(Agent):
         self.q_values = Signal("Q")
         self.signals.append(self.q_values)
 
+        self.reset_game(do_not_reset_env=True)
+
     # Algorithms for which q_values are calculated from predictions will override this function
     def get_q_values(self, prediction):
         return prediction