update.

doc/paddle/api/paddle/fluid/layers/nn/py_func_cn.rst

@@ -107,8 +107,8 @@ PaddlePaddle 通过py_func在Python端注册OP。py_func的设计原理在于Pad
         main_program = paddle.static.default_main_program()
 
         # 创建前向函数的输入变量
-        x = paddle.data(name='x', shape=[2,3], dtype='int32')
-        y = paddle.data(name='y', shape=[2,3], dtype='int32')
+        x = paddle.static.data(name='x', shape=[2,3], dtype='int32')
+        y = paddle.static.data(name='y', shape=[2,3], dtype='int32')
         
         # 创建前向函数的输出变量，必须指明变量名称name/数据类型dtype/维度shape
         output = create_tmp_var('output','int32', [3,1])

doc/paddle/api/paddle/fluid/param_attr/ParamAttr_cn.rst

@@ -42,7 +42,7 @@ ParamAttr
                                     trainable=True)
    print(w_param_attrs.name) # "fc_weight"
    paddle.enable_static()
-   x = paddle.data(name='X', shape=[1], dtype='float32')
+   x = paddle.static.data(name='X', shape=[1], dtype='float32')
    y_predict = paddle.static.nn.fc(input=x, size=10, param_attr=w_param_attrs)