diff --git a/paddle/gserver/layers/ScaleShiftLayer.cpp b/paddle/gserver/layers/ScaleShiftLayer.cpp
index 4f5b1c6225a70b96ec266674437175c7d60fc87c..06dcb409f8c98c23a1116513f6bcbd2446b290f4 100644
--- a/paddle/gserver/layers/ScaleShiftLayer.cpp
+++ b/paddle/gserver/layers/ScaleShiftLayer.cpp
@@ -17,8 +17,9 @@ limitations under the License. */
 namespace paddle {
 
 /**
- * A layer does scaling and shifting to the input by appling a slope and
- * an intercept which are trainable to the input element-wise.
+ * A layer applies a slope and an intercept to the input element-wise for
+ * scaling and shifting. Noting that this layer is trainable which differs
+ * from the SlopeInterceptLayer.
  *
  * \f[
  *    y = wx + b
diff --git a/paddle/gserver/tests/test_LayerGrad.cpp b/paddle/gserver/tests/test_LayerGrad.cpp
index 65429ebada1bcf350fb37b1b561698b95b0f5182..dd2c955e6a4660a1811f205ec5c5861798291912 100644
--- a/paddle/gserver/tests/test_LayerGrad.cpp
+++ b/paddle/gserver/tests/test_LayerGrad.cpp
@@ -2008,8 +2008,8 @@ TEST(Layer, RowL2NormLayer) {
 }
 
 TEST(Layer, ScaleShiftLayer) {
-  const size_t batchSize = 128;
-  const size_t size = 512;
+  const size_t batchSize = 16;
+  const size_t size = 32;
   TestConfig config;
   config.layerConfig.set_type("scale_shift");
   config.layerConfig.set_size(size);
diff --git a/python/paddle/trainer_config_helpers/layers.py b/python/paddle/trainer_config_helpers/layers.py
index 4c7217024ab2eb78f1adabd818f8e05a8d2a2091..ec3a87aa3663e4926554c16933116defc6724534 100755
--- a/python/paddle/trainer_config_helpers/layers.py
+++ b/python/paddle/trainer_config_helpers/layers.py
@@ -6219,8 +6219,9 @@ def kmax_sequence_score_layer(input, name=None, beam_size=1):
 @wrap_bias_attr_default()
 def scale_shift_layer(input, name=None, param_attr=None, bias_attr=None):
     """
-    A layer does scaling and shifting to the input by appling a slope and 
-    an intercept which are trainable to the input element-wise.
+    A layer applies a slope and an intercept to the input element-wise for 
+    scaling and shifting. Noting that this layer is trainable which differs
+    from the slope_intercept_layer.
     .. math::
 
         y = w * x + b
diff --git a/python/paddle/trainer_config_helpers/tests/configs/protostr/test_scale_shift_layer.protostr b/python/paddle/trainer_config_helpers/tests/configs/protostr/test_scale_shift_layer.protostr
index efaf20f8a76da3c92c05238898a28e16e915f249..35ade126a2586a8e3eee6f0ac3c7e49523c8f5c5 100644
--- a/python/paddle/trainer_config_helpers/tests/configs/protostr/test_scale_shift_layer.protostr
+++ b/python/paddle/trainer_config_helpers/tests/configs/protostr/test_scale_shift_layer.protostr
@@ -14,7 +14,6 @@ layers {
     input_layer_name: "data"
     input_parameter_name: "___scale_shift_0__.w0"
   }
-  bias_parameter_name: "___scale_shift_0__.wbias"
 }
 layers {
   name: "__scale_shift_1__"
@@ -25,6 +24,7 @@ layers {
     input_layer_name: "data"
     input_parameter_name: "___scale_shift_1__.w0"
   }
+  bias_parameter_name: "___scale_shift_1__.wbias"
 }
 parameters {
   name: "___scale_shift_0__.w0"
@@ -37,24 +37,24 @@ parameters {
   initial_smart: true
 }
 parameters {
-  name: "___scale_shift_0__.wbias"
+  name: "___scale_shift_1__.w0"
   size: 1
   initial_mean: 0.0
-  initial_std: 0.0
+  initial_std: 1.0
   dims: 1
   dims: 1
   initial_strategy: 0
-  initial_smart: false
+  initial_smart: true
 }
 parameters {
-  name: "___scale_shift_1__.w0"
+  name: "___scale_shift_1__.wbias"
   size: 1
   initial_mean: 0.0
-  initial_std: 1.0
+  initial_std: 0.0
   dims: 1
   dims: 1
   initial_strategy: 0
-  initial_smart: true
+  initial_smart: false
 }
 input_layer_names: "data"
 output_layer_names: "__scale_shift_0__"
diff --git a/python/paddle/trainer_config_helpers/tests/configs/test_scale_shift_layer.py b/python/paddle/trainer_config_helpers/tests/configs/test_scale_shift_layer.py
index 818d71f15dbe66b364c15beb334d0c09703bab1c..dd589116fa9932144ca066d3fa4c929d1433a7f1 100644
--- a/python/paddle/trainer_config_helpers/tests/configs/test_scale_shift_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_scale_shift_layer.py
@@ -1,11 +1,9 @@
 from paddle.trainer_config_helpers import *
 
-settings(batch_size=1000, learning_rate=1e-5)
-
 data = data_layer(name='data', size=100)
 
-scale = scale_shift_layer(input=data)
+scale = scale_shift_layer(input=data, bias_attr=False)
 
-scale_shift = scale_shift_layer(input=data, bias_attr=False)
+scale_shift = scale_shift_layer(input=data)
 
 outputs(scale, scale_shift)