diff --git a/parl/algorithms/fluid/a3c.py b/parl/algorithms/fluid/a3c.py
index 9b9f57e8eb5bfd59e3f79c1fc42e4d1374618f23..27aa1dc8785315e4347cebcc3b13f8d80659e0a9 100644
--- a/parl/algorithms/fluid/a3c.py
+++ b/parl/algorithms/fluid/a3c.py
@@ -72,6 +72,7 @@ class A3C(Algorithm):
         policy_entropy = policy_distribution.entropy()
         entropy = layers.reduce_sum(policy_entropy)
 
+        entropy_coeff = layers.reshape(entropy_coeff, shape=[1])
         total_loss = (
             pi_loss + vf_loss * self.vf_loss_coeff + entropy * entropy_coeff)
 
diff --git a/parl/algorithms/fluid/maddpg.py b/parl/algorithms/fluid/maddpg.py
index 4bf799413165d81d00238a7c156511a03619ba5d..36b14709aaf5e5e0a2cacc97bd94b1097caf2404 100644
--- a/parl/algorithms/fluid/maddpg.py
+++ b/parl/algorithms/fluid/maddpg.py
@@ -27,10 +27,11 @@ from parl.core.fluid.policy_distribution import SoftMultiCategoricalDistribution
 
 
 def SoftPDistribution(logits, act_space):
-    """input:
+    """Args:
             logits: the output of policy model
             act_space: action space, must be gym.spaces.Discrete or multiagent.multi_discrete.MultiDiscrete
-        output：
+
+        Return:
             instance of SoftCategoricalDistribution or SoftMultiCategoricalDistribution
     """
     # is instance of gym.spaces.Discrete