Merge pull request #261 from kuni-kuni/ppo_batch_states

Enable to change batch_states in PPO

chainerrl/agents/ppo.py

@@ -71,6 +71,7 @@ def __init__(self, model, optimizer,
                  clip_eps_vf=None,
                  standardize_advantages=True,
                  average_v_decay=0.999, average_loss_decay=0.99,
+                 batch_states=batch_states,
                  ):
         self.model = model
 
@@ -98,6 +99,8 @@ def __init__(self, model, optimizer,
         self.average_loss_entropy = 0
         self.average_loss_decay = average_loss_decay
 
+        self.batch_states = batch_states
+
         self.xp = self.model.xp
         self.last_state = None
 
@@ -107,7 +110,7 @@ def __init__(self, model, optimizer,
     def _act(self, state):
         xp = self.xp
         with chainer.using_config('train', False):
-            b_state = batch_states([state], xp, self.phi)
+            b_state = self.batch_states([state], xp, self.phi)
             with chainer.no_backprop_mode():
                 action_distrib, v = self.model(b_state)
                 action = action_distrib.sample()
@@ -200,7 +203,7 @@ def update(self):
         dataset_iter.reset()
         while dataset_iter.epoch < self.epochs:
             batch = dataset_iter.__next__()
-            states = batch_states([b['state'] for b in batch], xp, self.phi)
+            states = self.batch_states([b['state'] for b in batch], xp, self.phi)
             actions = xp.array([b['action'] for b in batch])
             distribs, vs_pred = self.model(states)
             with chainer.no_backprop_mode():
@@ -224,7 +227,7 @@ def update(self):
     def act_and_train(self, obs, reward):
         if hasattr(self.model, 'obs_filter'):
             xp = self.xp
-            b_state = batch_states([obs], xp, self.phi)
+            b_state = self.batch_states([obs], xp, self.phi)
             self.model.obs_filter.experience(b_state)
 
         action, v = self._act(obs)