implement dqn loss and dpg loss, add TODO for separate actor and critic

examples/dqn_example.py

@@ -53,7 +53,7 @@ if __name__ == '__main__':
     action = tf.placeholder(dtype=tf.int32, shape=[None]) # batch of integer actions
     target = tf.placeholder(dtype=tf.float32, shape=[None]) # target value for DQN
 
-    dqn_loss = losses.dqn_loss(action, target, pi) # TongzhengRen
+    dqn_loss = losses.dqn_loss(action, target, q_net) # TongzhengRen
 
     total_loss = dqn_loss
     optimizer = tf.train.AdamOptimizer(1e-3)

tianshou/core/README.md

@@ -1,3 +1,7 @@
+#TODO: 
+
+Separate actor and critic. (Important, we need to focus on that recently)
+
 # policy
 
 YongRen

tianshou/core/losses.py

@@ -26,7 +26,7 @@ def vanilla_policy_gradient(sampled_action, reward, pi, baseline="None"):
 
     :param sampled_action: placeholder of sampled actions during interaction with the environment
     :param reward: placeholder of reward the 'sampled_action' get
-    :param pi: current 'policy' to be optimized
+    :param pi: current `policy` to be optimized
     :param baseline: the baseline method used to reduce the variance, default is 'None'
     :return:
     """
@@ -35,8 +35,25 @@ def vanilla_policy_gradient(sampled_action, reward, pi, baseline="None"):
     # TODO： Different baseline methods like REINFORCE, etc.
     return vanilla_policy_gradient_loss
 
-def temporal_difference_loss():
-    pass
+def dqn_loss(sampled_action, sampled_target, q_net):
+    """
+    deep q-network
 
-def deterministic_policy_gradient():
-    pass
+    :param sampled_action: placeholder of sampled actions during the interaction with the environment
+    :param sampled_target: estimated Q(s,a)
+    :param q_net: current `policy` to be optimized
+    :return:
+    """
+    action_num = q_net.get_values().shape()[1]
+    sampled_q = tf.reduce_sum(q_net.get_values() * tf.one_hot(sampled_action, action_num), axis=1)
+    return tf.reduce_mean(tf.square(sampled_target - sampled_q))
+
+def deterministic_policy_gradient(sampled_state, critic):
+    """
+    deterministic policy gradient:
+
+    :param sampled_action: placeholder of sampled actions during the interaction with the environment
+    :param critic: current `value` function
+    :return:
+    """
+    return tf.reduce_mean(critic.get_value(sampled_state))

tianshou/core/policy/base.py

@@ -14,12 +14,14 @@ __all__ = [
     'StochasticPolicy',
 ]
 
+#TODO: separate actor and critic, we should focus on it once we finish the basic module.
+
 class QValuePolicy(object):
     """
     The policy as in DQN
     """
-    def __init__(self, value_tensor):
-        pass
+    def __init__(self, observation_placeholder):
+        self.observation_placeholder = observation_placeholder
 
     def act(self, observation, exploration=None): # first implement no exploration
         """
@@ -222,7 +224,3 @@ class StochasticPolicy(object):
         Private method for subclasses to rewrite the :meth:`prob` method.
         """
         raise NotImplementedError()
-
-
-class QValuePolicy(object):
-    pass

tianshou/data/README.md

@@ -1,3 +1,7 @@
+# TODO:
+
+Notice that we will separate actor and critic, and batch will collect data for optimizing policy while replay will collect data for optimizing critic.
+
 # Batch
 
 YouQiaoben