diff --git a/01.fit_a_line/README.cn.md b/01.fit_a_line/README.cn.md
index 715bc98e5dcbbcfb340478c5de8d448fa4d4cc55..69cc494df6608d9e1026efbe618e532b18fba8f3 100644
--- a/01.fit_a_line/README.cn.md
+++ b/01.fit_a_line/README.cn.md
@@ -194,8 +194,8 @@ test_reader = paddle.batch(
 训练程序的目的是定义一个训练模型的网络结构。对于线性回归来讲，它就是一个从输入到输出的简单的全连接层。更加复杂的结果，比如卷积神经网络，递归神经网络等会在随后的章节中介绍。训练程序必须返回`平均损失`作为第一个返回值，因为它会被后面反向传播算法所用到。
 
 ```python
-x = fluid.data(name='x', shape=[-1, 13], dtype='float32') # 定义输入的形状和数据类型
-y = fluid.data(name='y', shape=[-1, 1], dtype='float32') # 定义输出的形状和数据类型
+x = fluid.data(name='x', shape=[None, 13], dtype='float32') # 定义输入的形状和数据类型
+y = fluid.data(name='y', shape=[None, 1], dtype='float32') # 定义输出的形状和数据类型
 y_predict = fluid.layers.fc(input=x, size=1, act=None) # 连接输入和输出的全连接层
 
 main_program = fluid.default_main_program() # 获取默认/全局主函数
diff --git a/01.fit_a_line/README.md b/01.fit_a_line/README.md
index 29fdc1a6c72c4f3da0939688359fb98b296f87f6..8760bbf0c6ac488f183a3adedb7bf6cde4058942 100644
--- a/01.fit_a_line/README.md
+++ b/01.fit_a_line/README.md
@@ -196,8 +196,8 @@ test_reader = paddle.batch(
 The aim of the program for training is to define a network structure of a training model. For linear regression, it is a simple fully connected layer from input to output. More complex result, such as Convolutional Neural Network and Recurrent Neural Network, will be introduced in later chapters. It must return `mean error` as the first return value in program for training, for that `mean error` will be used for BackPropagation.
 
 ```python
-x = fluid.data(name='x', shape=[-1, 13], dtype='float32') # define shape and data type of input
-y = fluid.data(name='y', shape=[-1, 1], dtype='float32') # define shape and data type of output
+x = fluid.data(name='x', shape=[None, 13], dtype='float32') # define shape and data type of input
+y = fluid.data(name='y', shape=[None, 1], dtype='float32') # define shape and data type of output
 y_predict = fluid.layers.fc(input=x, size=1, act=None) # fully connected layer connecting input and output
 
 main_program = fluid.default_main_program() # get default/global main function
diff --git a/01.fit_a_line/index.cn.html b/01.fit_a_line/index.cn.html
index 40d91286ee06e375b056bdea95e5a2bbf5a43328..e47cda5f7e74d4e41d0d341fab126712dc78434a 100644
--- a/01.fit_a_line/index.cn.html
+++ b/01.fit_a_line/index.cn.html
@@ -236,8 +236,8 @@ test_reader = paddle.batch(
 训练程序的目的是定义一个训练模型的网络结构。对于线性回归来讲，它就是一个从输入到输出的简单的全连接层。更加复杂的结果，比如卷积神经网络，递归神经网络等会在随后的章节中介绍。训练程序必须返回`平均损失`作为第一个返回值，因为它会被后面反向传播算法所用到。
 
 ```python
-x = fluid.data(name='x', shape=[-1, 13], dtype='float32') # 定义输入的形状和数据类型
-y = fluid.data(name='y', shape=[-1, 1], dtype='float32') # 定义输出的形状和数据类型
+x = fluid.data(name='x', shape=[None, 13], dtype='float32') # 定义输入的形状和数据类型
+y = fluid.data(name='y', shape=[None, 1], dtype='float32') # 定义输出的形状和数据类型
 y_predict = fluid.layers.fc(input=x, size=1, act=None) # 连接输入和输出的全连接层
 
 main_program = fluid.default_main_program() # 获取默认/全局主函数
diff --git a/01.fit_a_line/index.html b/01.fit_a_line/index.html
index 184e8253bda3711afe6f3f2d60fc21aab2bcdb1c..70e7b0eab98652434f4e030e7888d07290176ef3 100644
--- a/01.fit_a_line/index.html
+++ b/01.fit_a_line/index.html
@@ -238,8 +238,8 @@ test_reader = paddle.batch(
 The aim of the program for training is to define a network structure of a training model. For linear regression, it is a simple fully connected layer from input to output. More complex result, such as Convolutional Neural Network and Recurrent Neural Network, will be introduced in later chapters. It must return `mean error` as the first return value in program for training, for that `mean error` will be used for BackPropagation.
 
 ```python
-x = fluid.data(name='x', shape=[-1, 13], dtype='float32') # define shape and data type of input
-y = fluid.data(name='y', shape=[-1, 1], dtype='float32') # define shape and data type of output
+x = fluid.data(name='x', shape=[None, 13], dtype='float32') # define shape and data type of input
+y = fluid.data(name='y', shape=[None, 1], dtype='float32') # define shape and data type of output
 y_predict = fluid.layers.fc(input=x, size=1, act=None) # fully connected layer connecting input and output
 
 main_program = fluid.default_main_program() # get default/global main function
diff --git a/01.fit_a_line/train.py b/01.fit_a_line/train.py
index 2ebe21e6fd9319c18369d177684ed62186bedd82..e7088986edace4df0a795c768ddc50ad0d9dde74 100644
--- a/01.fit_a_line/train.py
+++ b/01.fit_a_line/train.py
@@ -87,8 +87,8 @@ def main():
             batch_size=batch_size)
 
     # feature vector of length 13
-    x = fluid.data(name='x', shape=[-1, 13], dtype='float32')
-    y = fluid.data(name='y', shape=[-1, 1], dtype='float32')
+    x = fluid.data(name='x', shape=[None, 13], dtype='float32')
+    y = fluid.data(name='y', shape=[None, 1], dtype='float32')
 
     main_program = fluid.default_main_program()
     startup_program = fluid.default_startup_program()
diff --git a/02.recognize_digits/README.cn.md b/02.recognize_digits/README.cn.md
index 31c2adbb6b27fd53179643af099f336d713d3483..b36211907eafcc12d418b2eb380de10f8cf512c8 100644
--- a/02.recognize_digits/README.cn.md
+++ b/02.recognize_digits/README.cn.md
@@ -209,7 +209,7 @@ def softmax_regression():
         predict_image -- 分类的结果
     """
     # 输入的原始图像数据，大小为28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # 以softmax为激活函数的全连接层，输出层的大小必须为数字的个数10
     predict = fluid.layers.fc(
         input=img, size=10, act='softmax')
@@ -229,7 +229,7 @@ def multilayer_perceptron():
         predict_image -- 分类的结果
     """
     # 输入的原始图像数据，大小为28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # 第一个全连接层，激活函数为ReLU
     hidden = fluid.layers.fc(input=img, size=200, act='relu')
     # 第二个全连接层，激活函数为ReLU
@@ -251,7 +251,7 @@ def convolutional_neural_network():
         predict -- 分类的结果
     """
     # 输入的原始图像数据，大小为28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # 第一个卷积-池化层
     # 使用20个5*5的滤波器，池化大小为2，池化步长为2，激活函数为Relu
     conv_pool_1 = fluid.nets.simple_img_conv_pool(
@@ -296,7 +296,7 @@ def train_program():
 
     """
     # 标签层，名称为label,对应输入图片的类别标签
-    label = fluid.data(name='label', shape=[-1, 1], dtype='int64')
+    label = fluid.data(name='label', shape=[None, 1], dtype='int64')
 
     # predict = softmax_regression() # 取消注释将使用 Softmax回归
     # predict = multilayer_perceptron() # 取消注释将使用 多层感知器
diff --git a/02.recognize_digits/README.md b/02.recognize_digits/README.md
index b3ca3c86f4b300841a6e0df7733453e0762c27a6..47691ef026b341e831cb8d828ad5799e3dd5264e 100644
--- a/02.recognize_digits/README.md
+++ b/02.recognize_digits/README.md
@@ -188,7 +188,7 @@ def softmax_regression():
     predict_image -- result of classification
     """
     # input original image data in size of 28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # With softmax as the fully connected layer of the activation function, the size of the output layer must be 10
     predict = fluid.layers.fc(
     input=img, size=10, act='softmax')
@@ -208,7 +208,7 @@ def multilayer_perceptron():
     predict_image -- result of classification
     """
     # input raw image data in size of 28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # the first fully connected layer, whose activation function is ReLU
     hidden = fluid.layers.fc(input=img, size=200, act='relu')
     # the second fully connected layer, whose activation function is ReLU
@@ -230,7 +230,7 @@ def convolutional_neural_network():
     predict -- result of classification
     """
     # input raw image data in size of 28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # the first convolution-pooling layer
     # Use 20 5*5 filters, the pooling size is 2, the pooling step is 2, and the activation function is Relu.
     conv_pool_1 = fluid.nets.simple_img_conv_pool(
@@ -275,7 +275,7 @@ def train_program():
 
     """
     # label layer, called label, correspondent with label category of input picture
-    label = fluid.data(name='label', shape=[-1, 1], dtype='int64')
+    label = fluid.data(name='label', shape=[None, 1], dtype='int64')
 
     # predict = softmax_regression() # cancel note and run Softmax regression
     # predict = multilayer_perceptron() # cancel note and run multiple perceptron
diff --git a/02.recognize_digits/index.cn.html b/02.recognize_digits/index.cn.html
index 328815220ae2521fdaf91e7bb4cf3ef3aa29698f..d3f82cd1a86bbfdb994213846dc1acf31e7fc7d3 100644
--- a/02.recognize_digits/index.cn.html
+++ b/02.recognize_digits/index.cn.html
@@ -251,7 +251,7 @@ def softmax_regression():
         predict_image -- 分类的结果
     """
     # 输入的原始图像数据，大小为28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # 以softmax为激活函数的全连接层，输出层的大小必须为数字的个数10
     predict = fluid.layers.fc(
         input=img, size=10, act='softmax')
@@ -271,7 +271,7 @@ def multilayer_perceptron():
         predict_image -- 分类的结果
     """
     # 输入的原始图像数据，大小为28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # 第一个全连接层，激活函数为ReLU
     hidden = fluid.layers.fc(input=img, size=200, act='relu')
     # 第二个全连接层，激活函数为ReLU
@@ -293,7 +293,7 @@ def convolutional_neural_network():
         predict -- 分类的结果
     """
     # 输入的原始图像数据，大小为28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # 第一个卷积-池化层
     # 使用20个5*5的滤波器，池化大小为2，池化步长为2，激活函数为Relu
     conv_pool_1 = fluid.nets.simple_img_conv_pool(
@@ -338,7 +338,7 @@ def train_program():
 
     """
     # 标签层，名称为label,对应输入图片的类别标签
-    label = fluid.data(name='label', shape=[-1, 1], dtype='int64')
+    label = fluid.data(name='label', shape=[None, 1], dtype='int64')
 
     # predict = softmax_regression() # 取消注释将使用 Softmax回归
     # predict = multilayer_perceptron() # 取消注释将使用 多层感知器
diff --git a/02.recognize_digits/index.html b/02.recognize_digits/index.html
index 9ab5b1eccc9c35a1364b486ede01571d780d7c4e..04ff588588e54ae953873470f60f7724a6856492 100644
--- a/02.recognize_digits/index.html
+++ b/02.recognize_digits/index.html
@@ -230,7 +230,7 @@ def softmax_regression():
     predict_image -- result of classification
     """
     # input original image data in size of 28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # With softmax as the fully connected layer of the activation function, the size of the output layer must be 10
     predict = fluid.layers.fc(
     input=img, size=10, act='softmax')
@@ -250,7 +250,7 @@ def multilayer_perceptron():
     predict_image -- result of classification
     """
     # input raw image data in size of 28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # the first fully connected layer, whose activation function is ReLU
     hidden = fluid.layers.fc(input=img, size=200, act='relu')
     # the second fully connected layer, whose activation function is ReLU
@@ -272,7 +272,7 @@ def convolutional_neural_network():
     predict -- result of classification
     """
     # input raw image data in size of 28*28*1
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
     # the first convolution-pooling layer
     # Use 20 5*5 filters, the pooling size is 2, the pooling step is 2, and the activation function is Relu.
     conv_pool_1 = fluid.nets.simple_img_conv_pool(
@@ -317,7 +317,7 @@ def train_program():
 
     """
     # label layer, called label, correspondent with label category of input picture
-    label = fluid.data(name='label', shape=[-1, 1], dtype='int64')
+    label = fluid.data(name='label', shape=[None, 1], dtype='int64')
 
     # predict = softmax_regression() # cancel note and run Softmax regression
     # predict = multilayer_perceptron() # cancel note and run multiple perceptron
diff --git a/02.recognize_digits/train.py b/02.recognize_digits/train.py
index 27cb3c72ed20a63993c5fc1c11cd222e5d72b5ef..1d1fa9a7c64e9c9dd99f0ae3b0234a139b4d6de0 100644
--- a/02.recognize_digits/train.py
+++ b/02.recognize_digits/train.py
@@ -101,8 +101,8 @@ def train(nn_type,
         test_reader = paddle.batch(
             paddle.dataset.mnist.test(), batch_size=BATCH_SIZE)
 
-    img = fluid.data(name='img', shape=[-1, 1, 28, 28], dtype='float32')
-    label = fluid.data(name='label', shape=[-1, 1], dtype='int64')
+    img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
+    label = fluid.data(name='label', shape=[None, 1], dtype='int64')
 
     if nn_type == 'softmax_regression':
         net_conf = softmax_regression
diff --git a/09.gan/README.cn.md b/09.gan/README.cn.md
index 99cf01ed570e122248a38982eeaf5b284034226a..4c0fd2298b34f8fcb87202039a9bf1d9542d279e 100644
--- a/09.gan/README.cn.md
+++ b/09.gan/README.cn.md
@@ -265,16 +265,16 @@ dg_program = fluid.Program()
 # 定义判别真实图片的program
 with fluid.program_guard(d_program):
     # 输入图片大小为28*28=784
-    img = fluid.layers.data(name='img', shape=[784], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 784], dtype='float32')
     # 标签shape=1
-    label = fluid.layers.data(name='label', shape=[1], dtype='float32')
+    label = fluid.data(name='label', shape=[None, 1], dtype='float32')
     d_logit = D(img)
     d_loss = loss(d_logit, label)
 
 # 定义判别生成图片的program
 with fluid.program_guard(dg_program):
-    noise = fluid.layers.data(
-        name='noise', shape=[NOISE_SIZE], dtype='float32')
+    noise = fluid.data(
+        name='noise', shape=[None, NOISE_SIZE], dtype='float32')
     # 噪声数据作为输入得到生成图片
     g_img = G(x=noise)
 
diff --git a/09.gan/dc_gan.py b/09.gan/dc_gan.py
index c08752aecf886c1fb882604b5116f3ec9ea1f00d..7234dc837109a1132612cff759634a8330c3c1e9 100644
--- a/09.gan/dc_gan.py
+++ b/09.gan/dc_gan.py
@@ -60,14 +60,14 @@ def train(args):
     dg_program = fluid.Program()
 
     with fluid.program_guard(d_program):
-        img = fluid.layers.data(name='img', shape=[784], dtype='float32')
-        label = fluid.layers.data(name='label', shape=[1], dtype='float32')
+        img = fluid.data(name='img', shape=[None, 784], dtype='float32')
+        label = fluid.data(name='label', shape=[None, 1], dtype='float32')
         d_logit = D(img)
         d_loss = loss(d_logit, label)
 
     with fluid.program_guard(dg_program):
-        noise = fluid.layers.data(
-            name='noise', shape=[NOISE_SIZE], dtype='float32')
+        noise = fluid.data(
+            name='noise', shape=[None, NOISE_SIZE], dtype='float32')
         g_img = G(x=noise)
 
         g_program = dg_program.clone()
diff --git a/09.gan/index.cn.html b/09.gan/index.cn.html
index 8947f5a1725538fea260a17b89f57e6f028a2336..a6b88e8ae0ad5a09bacae26410dba829a55c0c05 100644
--- a/09.gan/index.cn.html
+++ b/09.gan/index.cn.html
@@ -307,16 +307,16 @@ dg_program = fluid.Program()
 # 定义判别真实图片的program
 with fluid.program_guard(d_program):
     # 输入图片大小为28*28=784
-    img = fluid.layers.data(name='img', shape=[784], dtype='float32')
+    img = fluid.data(name='img', shape=[None, 784], dtype='float32')
     # 标签shape=1
-    label = fluid.layers.data(name='label', shape=[1], dtype='float32')
+    label = fluid.data(name='label', shape=[None, 1], dtype='float32')
     d_logit = D(img)
     d_loss = loss(d_logit, label)
 
 # 定义判别生成图片的program
 with fluid.program_guard(dg_program):
-    noise = fluid.layers.data(
-        name='noise', shape=[NOISE_SIZE], dtype='float32')
+    noise = fluid.data(
+        name='noise', shape=[None, NOISE_SIZE], dtype='float32')
     # 噪声数据作为输入得到生成图片
     g_img = G(x=noise)