diff --git a/fluid/PaddleCV/image_classification/models/alexnet.py b/fluid/PaddleCV/image_classification/models/alexnet.py
index 3e0eab2dee1d2f2e8d3cb2e8c12a3504a1e7c0e5..abe3b92965b1c16312e4ddf68809f6a4c93183fa 100644
--- a/fluid/PaddleCV/image_classification/models/alexnet.py
+++ b/fluid/PaddleCV/image_classification/models/alexnet.py
@@ -142,7 +142,6 @@ class AlexNet():
         out = fluid.layers.fc(
             input=fc7,
             size=class_dim,
-            act='softmax',
             bias_attr=fluid.param_attr.ParamAttr(
                 initializer=fluid.initializer.Uniform(-stdv, stdv)),
             param_attr=fluid.param_attr.ParamAttr(
diff --git a/fluid/PaddleCV/image_classification/models/dpn.py b/fluid/PaddleCV/image_classification/models/dpn.py
index ca49898fb76d10344e0b847b3631b194192d7e5e..316e96ac2cebd2dec4a60bf1748ed321fa651590 100644
--- a/fluid/PaddleCV/image_classification/models/dpn.py
+++ b/fluid/PaddleCV/image_classification/models/dpn.py
@@ -94,7 +94,6 @@ class DPN(object):
             initializer=fluid.initializer.Uniform(-stdv, stdv))
         fc6 = fluid.layers.fc(input=pool5,
                               size=class_dim,
-                              act='softmax',
                               param_attr=param_attr)
 
         return fc6
diff --git a/fluid/PaddleCV/image_classification/models/inception_v4.py b/fluid/PaddleCV/image_classification/models/inception_v4.py
index d3a80a20500f365166d50a0cf222613d0427354f..1520375477ade6e61f0a5584278b13e40ab541eb 100644
--- a/fluid/PaddleCV/image_classification/models/inception_v4.py
+++ b/fluid/PaddleCV/image_classification/models/inception_v4.py
@@ -47,7 +47,6 @@ class InceptionV4():
         out = fluid.layers.fc(
             input=drop,
             size=class_dim,
-            act='softmax',
             param_attr=fluid.param_attr.ParamAttr(
                 initializer=fluid.initializer.Uniform(-stdv, stdv)))
         return out
diff --git a/fluid/PaddleCV/image_classification/models/mobilenet.py b/fluid/PaddleCV/image_classification/models/mobilenet.py
index f3554734768d5bbec96dac2443b48389d235da91..d0b419e8b4083104ba529c9f886284aa724953e6 100644
--- a/fluid/PaddleCV/image_classification/models/mobilenet.py
+++ b/fluid/PaddleCV/image_classification/models/mobilenet.py
@@ -120,7 +120,6 @@ class MobileNet():
 
         output = fluid.layers.fc(input=input,
                                  size=class_dim,
-                                 act='softmax',
                                  param_attr=ParamAttr(initializer=MSRA()))
         return output
 
diff --git a/fluid/PaddleCV/image_classification/models/mobilenet_v2.py b/fluid/PaddleCV/image_classification/models/mobilenet_v2.py
index 9ec118f2605e5aa4815668add1d1e2b3dae7cf7c..c219b1bf5a7260fbb07627bc3fa039f4b2833092 100644
--- a/fluid/PaddleCV/image_classification/models/mobilenet_v2.py
+++ b/fluid/PaddleCV/image_classification/models/mobilenet_v2.py
@@ -73,7 +73,6 @@ class MobileNetV2():
 
         output = fluid.layers.fc(input=input,
                                  size=class_dim,
-                                 act='softmax',
                                  param_attr=ParamAttr(initializer=MSRA()))
         return output
 
diff --git a/fluid/PaddleCV/image_classification/models/resnet.py b/fluid/PaddleCV/image_classification/models/resnet.py
index 75c7b750541c60821a624f95a2ab56d2890fcb25..def99db6d84673b77582cf93374f4cb2f00e9ac5 100644
--- a/fluid/PaddleCV/image_classification/models/resnet.py
+++ b/fluid/PaddleCV/image_classification/models/resnet.py
@@ -60,7 +60,6 @@ class ResNet():
         stdv = 1.0 / math.sqrt(pool.shape[1] * 1.0)
         out = fluid.layers.fc(input=pool,
                               size=class_dim,
-                              act='softmax',
                               param_attr=fluid.param_attr.ParamAttr(
                                   initializer=fluid.initializer.Uniform(-stdv,
                                                                         stdv)))
diff --git a/fluid/PaddleCV/image_classification/models/resnet_dist.py b/fluid/PaddleCV/image_classification/models/resnet_dist.py
index 2dab3e6111d8d02df44233c0440a8ea9ce74faa5..9aed8a9841d4ae9d47cbbe15df51ab8652d5e3fc 100644
--- a/fluid/PaddleCV/image_classification/models/resnet_dist.py
+++ b/fluid/PaddleCV/image_classification/models/resnet_dist.py
@@ -62,7 +62,6 @@ class DistResNet():
         stdv = 1.0 / math.sqrt(pool.shape[1] * 1.0)
         out = fluid.layers.fc(input=pool,
                               size=class_dim,
-                              act='softmax',
                               param_attr=fluid.param_attr.ParamAttr(
                                   initializer=fluid.initializer.Uniform(-stdv,
                                                                         stdv),
diff --git a/fluid/PaddleCV/image_classification/models/se_resnext.py b/fluid/PaddleCV/image_classification/models/se_resnext.py
index 023046a5db506d0111e246b03810b02c70f3e1b8..ac50bd87b5070000a018949e777a897427c3e5a5 100644
--- a/fluid/PaddleCV/image_classification/models/se_resnext.py
+++ b/fluid/PaddleCV/image_classification/models/se_resnext.py
@@ -110,7 +110,6 @@ class SE_ResNeXt():
         stdv = 1.0 / math.sqrt(drop.shape[1] * 1.0)
         out = fluid.layers.fc(input=drop,
                               size=class_dim,
-                              act='softmax',
                               param_attr=fluid.param_attr.ParamAttr(
                                   initializer=fluid.initializer.Uniform(-stdv,
                                                                         stdv)))
diff --git a/fluid/PaddleCV/image_classification/models/shufflenet_v2.py b/fluid/PaddleCV/image_classification/models/shufflenet_v2.py
index bc61508e57fe0e8aec743b4a389c75bac5b3bd44..6db88aa769dd6b3b3e2987fcac6d8054319a2a56 100644
--- a/fluid/PaddleCV/image_classification/models/shufflenet_v2.py
+++ b/fluid/PaddleCV/image_classification/models/shufflenet_v2.py
@@ -93,7 +93,6 @@ class ShuffleNetV2():
 
         output = fluid.layers.fc(input=pool_last,
                                  size=class_dim,
-                                 act='softmax',
                                  param_attr=ParamAttr(initializer=MSRA()))
         return output
 
diff --git a/fluid/PaddleCV/image_classification/models/vgg.py b/fluid/PaddleCV/image_classification/models/vgg.py
index 1af664fdb554f05ba8a556abbba36cd1c3141a40..7f559982334575c7c2bc778e1be8a4ebf69549fc 100644
--- a/fluid/PaddleCV/image_classification/models/vgg.py
+++ b/fluid/PaddleCV/image_classification/models/vgg.py
@@ -64,7 +64,6 @@ class VGGNet():
         out = fluid.layers.fc(
             input=fc2,
             size=class_dim,
-            act='softmax',
             param_attr=fluid.param_attr.ParamAttr(
                 initializer=fluid.initializer.Normal(scale=0.005)),
             bias_attr=fluid.param_attr.ParamAttr(
diff --git a/fluid/PaddleCV/image_classification/models_name/alexnet.py b/fluid/PaddleCV/image_classification/models_name/alexnet.py
index 3dfaa25ff790b3952acb304118e6489bb2b58844..f063c4d6deb88905aaa5f8a5eba59903f58293e8 100644
--- a/fluid/PaddleCV/image_classification/models_name/alexnet.py
+++ b/fluid/PaddleCV/image_classification/models_name/alexnet.py
@@ -159,7 +159,6 @@ class AlexNet():
         out = fluid.layers.fc(
             input=fc7,
             size=class_dim,
-            act='softmax',
             bias_attr=fluid.param_attr.ParamAttr(
                 initializer=fluid.initializer.Uniform(-stdv, stdv),
                 name=layer_name[7] + "_offset"),
diff --git a/fluid/PaddleCV/image_classification/models_name/dpn.py b/fluid/PaddleCV/image_classification/models_name/dpn.py
index a39d907205cedf86848df9c950202c640fdd9dcb..7f759b3bb6bfa9c866e129ac93ab2c6a9cf4168c 100644
--- a/fluid/PaddleCV/image_classification/models_name/dpn.py
+++ b/fluid/PaddleCV/image_classification/models_name/dpn.py
@@ -122,7 +122,6 @@ class DPN(object):
             initializer=fluid.initializer.Uniform(-stdv, stdv))
         fc6 = fluid.layers.fc(input=pool5,
                               size=class_dim,
-                              act='softmax',
                               param_attr=param_attr,
                               name="fc6")
 
diff --git a/fluid/PaddleCV/image_classification/models_name/inception_v4.py b/fluid/PaddleCV/image_classification/models_name/inception_v4.py
index 7b857d456bad99b98d2520653050659ef6c58983..8c6c0dbb129f903b4f0b849f930a520b5f17e5db 100644
--- a/fluid/PaddleCV/image_classification/models_name/inception_v4.py
+++ b/fluid/PaddleCV/image_classification/models_name/inception_v4.py
@@ -48,7 +48,6 @@ class InceptionV4():
         out = fluid.layers.fc(
             input=drop,
             size=class_dim,
-            act='softmax',
             param_attr=ParamAttr(
                 initializer=fluid.initializer.Uniform(-stdv, stdv),
                 name="final_fc_weights"),
diff --git a/fluid/PaddleCV/image_classification/models_name/mobilenet.py b/fluid/PaddleCV/image_classification/models_name/mobilenet.py
index 2ac6b46abdad13ad5e5507605cd7e1ce9861f620..d242bc946a7b4bec9c9d2e34da2496c0901ba870 100644
--- a/fluid/PaddleCV/image_classification/models_name/mobilenet.py
+++ b/fluid/PaddleCV/image_classification/models_name/mobilenet.py
@@ -130,7 +130,6 @@ class MobileNet():
 
         output = fluid.layers.fc(input=input,
                                  size=class_dim,
-                                 act='softmax',
                                  param_attr=ParamAttr(
                                      initializer=MSRA(), name="fc7_weights"),
                                  bias_attr=ParamAttr(name="fc7_offset"))
diff --git a/fluid/PaddleCV/image_classification/models_name/mobilenet_v2.py b/fluid/PaddleCV/image_classification/models_name/mobilenet_v2.py
index 442bd67f212d7a3b6c1a0b900ba1489250d670af..77d88c7da625c0c953c75d229148868f0481f2a2 100644
--- a/fluid/PaddleCV/image_classification/models_name/mobilenet_v2.py
+++ b/fluid/PaddleCV/image_classification/models_name/mobilenet_v2.py
@@ -80,7 +80,6 @@ class MobileNetV2():
 
         output = fluid.layers.fc(input=input,
                                  size=class_dim,
-                                 act='softmax',
                                  param_attr=ParamAttr(name='fc10_weights'),
                                  bias_attr=ParamAttr(name='fc10_offset'))
         return output
diff --git a/fluid/PaddleCV/image_classification/models_name/resnet.py b/fluid/PaddleCV/image_classification/models_name/resnet.py
index 095bd155b2c2b329e4ce89e94ae410c104a40fd1..19fa4ff2c4c30d0fa11b592c21f3db5e51663159 100644
--- a/fluid/PaddleCV/image_classification/models_name/resnet.py
+++ b/fluid/PaddleCV/image_classification/models_name/resnet.py
@@ -74,7 +74,6 @@ class ResNet():
         stdv = 1.0 / math.sqrt(pool.shape[1] * 1.0)
         out = fluid.layers.fc(input=pool,
                               size=class_dim,
-                              act='softmax',
                               param_attr=fluid.param_attr.ParamAttr(
                                   initializer=fluid.initializer.Uniform(-stdv,
                                                                         stdv)))
diff --git a/fluid/PaddleCV/image_classification/models_name/se_resnext.py b/fluid/PaddleCV/image_classification/models_name/se_resnext.py
index e083c524f8cefaa7d2483725994ebabae91101e4..0ae3d66fddbe2d1b9da5e2f52fe80d15931d256d 100644
--- a/fluid/PaddleCV/image_classification/models_name/se_resnext.py
+++ b/fluid/PaddleCV/image_classification/models_name/se_resnext.py
@@ -123,7 +123,6 @@ class SE_ResNeXt():
         out = fluid.layers.fc(
             input=drop,
             size=class_dim,
-            act='softmax',
             param_attr=ParamAttr(
                 initializer=fluid.initializer.Uniform(-stdv, stdv),
                 name='fc6_weights'),
diff --git a/fluid/PaddleCV/image_classification/models_name/shufflenet_v2.py b/fluid/PaddleCV/image_classification/models_name/shufflenet_v2.py
index 1cd4767ea40b31dd57f4ad3142f2a4533eba1aa9..595debf2199be9609100cb686aad65eb9cb55416 100644
--- a/fluid/PaddleCV/image_classification/models_name/shufflenet_v2.py
+++ b/fluid/PaddleCV/image_classification/models_name/shufflenet_v2.py
@@ -97,7 +97,6 @@ class ShuffleNetV2():
 
         output = fluid.layers.fc(input=pool_last,
                                  size=class_dim,
-                                 act='softmax',
                                  param_attr=ParamAttr(
                                      initializer=MSRA(), name='fc6_weights'),
                                  bias_attr=ParamAttr(name='fc6_offset'))
diff --git a/fluid/PaddleCV/image_classification/models_name/vgg.py b/fluid/PaddleCV/image_classification/models_name/vgg.py
index ac26791a6a3ec2aba8fe35ef46bbfd05a30c4777..8fcd2d9f1c397a428685cfb7bd264f18c0d0a7e7 100644
--- a/fluid/PaddleCV/image_classification/models_name/vgg.py
+++ b/fluid/PaddleCV/image_classification/models_name/vgg.py
@@ -61,7 +61,6 @@ class VGGNet():
         out = fluid.layers.fc(
             input=fc2,
             size=class_dim,
-            act='softmax',
             param_attr=fluid.param_attr.ParamAttr(name=fc_name[2] + "_weights"),
             bias_attr=fluid.param_attr.ParamAttr(name=fc_name[2] + "_offset"))