diff --git a/paddle/operators/fc_op.cc b/paddle/operators/fc_op.cc
index be0fca3c7189b5a5709a57805e0fc374ae9ca20f..14a7fa8467d17ef3709738a11816c84f92da343b 100644
--- a/paddle/operators/fc_op.cc
+++ b/paddle/operators/fc_op.cc
@@ -41,10 +41,19 @@ class FCOp : public NetOp {
                       "The size of inputs X(%d) should be no less than 1.", n);
 
     auto x_num_col_dims = Attr<std::vector<int>>("xNumColDims");
-    PADDLE_ENFORCE_EQ(x_num_col_dims.size(), n,
-                      "The size of attribute xNumColDims(%d) should be the "
-                      "same as that of inputs X(%d).",
-                      x_num_col_dims.size(), n);
+
+    // Set all values or set no values (use the default value)
+    if (!x_num_col_dims.empty()) {
+      PADDLE_ENFORCE_EQ(x_num_col_dims.size(), n,
+                        "The size of attribute xNumColDims(%d) should be the "
+                        "same as that of inputs X(%d).",
+                        x_num_col_dims.size(), n);
+    } else {
+      x_num_col_dims.resize(n);
+      for (size_t i = 0; i < n; i++) {
+        x_num_col_dims[i] = 1;
+      }
+    }
 
     // mul_out[i] = X[i] * W[i]
     for (size_t i = 0; i < n; i++) {
@@ -81,7 +90,7 @@ class FCOp : public NetOp {
 
     auto activation = Attr<std::string>("activation");
     AppendOp(framework::OpRegistry::CreateOp(
-        activation, {{"X", {Output(add_out)}}}, {{"Y", {Output("Y")}}}, {}));
+        activation, {{"X", {Output(add_out)}}}, {{"Y", {Output("Out")}}}, {}));
     CompleteAddOp(false);
   }
 };
@@ -105,7 +114,7 @@ class FCOpMaker : public framework::OpProtoAndCheckerMaker {
              "(Tensor) the bias of FC operator, a 1-D vector of size "
              "number_of_neurons.");
 
-    AddOutput("Y",
+    AddOutput("Out",
               "(Tensor) the activated output matrix of FC operator, a 2-D "
               "matrix of size (minibatch, number_of_neurons).");
     AddOutput("MulOut",
@@ -137,7 +146,8 @@ class FCOpMaker : public framework::OpProtoAndCheckerMaker {
         "`X_i.dims[0] x ... x X_i.dims[xNumColDims_i - 1]`. "
         "The matrix's second dimension (the length of row) will be the product "
         "of `X_i`'s first `rank - xNumColDims_i` dimensions, that is "
-        "`X_i.dims[xNumColDims_i] x ... x X_i.dims[rank - 1]`)");
+        "`X_i.dims[xNumColDims_i] x ... x X_i.dims[rank - 1]`)")
+        .SetDefault(std::vector<int>{});
 
     AddComment(R"DOC(
 Fully Connected Operator, known as Fully Connected Layer or Inner Product Layer
@@ -148,13 +158,13 @@ learned weights with a matrix multiplication followed by a bias offset
 (optionally).
 
 Equation:
-  Y = Act(sum_n{X_i * W_i} + B)
+  Out = Act(sum_n{X_i * W_i} + B)
 
 where X_i is Tensor that will be reshaped to a 2-D matrix of size (M x K),
 usually M is the minibatch size and K is the number of input features.
 W_i is a 2-D matrix of size (K x N), where N means the number of neurons
 in the fully connected layer. B is a 1-D vector of size N.
-Thus, the output Y is a 2-D matrix of size (M x N).
+Thus, the output Out is a 2-D matrix of size (M x N).
 Activation type can be set to `identity` (default), `sigmoid` or `softmax`.
 )DOC");
   }
diff --git a/python/paddle/v2/framework/tests/test_fc_op.py b/python/paddle/v2/framework/tests/test_fc_op.py
index f646fad33735549c5133c60bddb2e834986fc2f5..ed8d869a407298fb3291b8fcb4dff56c257ad097 100644
--- a/python/paddle/v2/framework/tests/test_fc_op.py
+++ b/python/paddle/v2/framework/tests/test_fc_op.py
@@ -20,15 +20,14 @@ class TestFCOp1(OpTest):
             "MulOut": [("MulOut0", mul_out0)],
             "SumOut": sum_out,
             "AddOut": add_out,
-            "Y": identity_out
+            "Out": identity_out
         }
-        self.attrs = {"xNumColDims": [1]}
 
     def test_check_output(self):
         self.check_output()
 
     def test_check_grad(self):
-        self.check_grad(["X0", "W0", "B"], "Y", max_relative_error=0.01)
+        self.check_grad(["X0", "W0", "B"], "Out", max_relative_error=0.01)
 
 
 class TestFCOp2(OpTest):
@@ -56,7 +55,7 @@ class TestFCOp2(OpTest):
             "MulOut": [("MulOut0", mul_out0), ("MulOut1", mul_out1)],
             "SumOut": sum_out,
             "AddOut": add_out,
-            "Y": sigmoid_out
+            "Out": sigmoid_out
         }
 
     def test_check_output(self):
@@ -64,7 +63,7 @@ class TestFCOp2(OpTest):
 
     def test_check_grad(self):
         self.check_grad(
-            ["X0", "X1", "W0", "W1", "B"], "Y", max_relative_error=0.01)
+            ["X0", "X1", "W0", "W1", "B"], "Out", max_relative_error=0.01)
 
 
 if __name__ == '__main__':