diff --git a/ctr/network_conf.py b/ctr/network_conf.py
index d577abcfcf3adf57f8278ae7f0bc7e48a3621a06..a90d1dc6e8da0b1379926dd7b37ac4cc0d408a2a 100644
--- a/ctr/network_conf.py
+++ b/ctr/network_conf.py
@@ -100,6 +100,6 @@ class CTRmodel(object):
         self.output = layer.fc(
             input=merge_layer, size=1, act=paddle.activation.Sigmoid())
         if not self.is_infer:
-            self.train_cost = paddle.layer.mse_cost(
+            self.train_cost = paddle.layer.square_error_cost(
                 input=self.output, label=self.click)
         return self.output
diff --git a/dssm/network_conf.py b/dssm/network_conf.py
index 996e91ae1e8daf0b7f52f7dae8dfdfef78dfabc7..04c2b7e2f389d3839b72d005004119a49238444f 100644
--- a/dssm/network_conf.py
+++ b/dssm/network_conf.py
@@ -270,7 +270,7 @@ class DSSM(object):
                 input=prediction, label=label)
         else:
             prediction = paddle.layer.cos_sim(*semantics)
-            cost = paddle.layer.mse_cost(prediction, label)
+            cost = paddle.layer.square_error_cost(prediction, label)
 
         if not self.is_infer:
             return cost, prediction, label