diff --git a/python/paddle/v2/data_set/__init__.py b/python/paddle/v2/dataset/__init__.py
similarity index 100%
rename from python/paddle/v2/data_set/__init__.py
rename to python/paddle/v2/dataset/__init__.py
diff --git a/python/paddle/v2/data_set/config.py b/python/paddle/v2/dataset/config.py
similarity index 100%
rename from python/paddle/v2/data_set/config.py
rename to python/paddle/v2/dataset/config.py
diff --git a/python/paddle/v2/data_set/mnist.py b/python/paddle/v2/dataset/mnist.py
similarity index 94%
rename from python/paddle/v2/data_set/mnist.py
rename to python/paddle/v2/dataset/mnist.py
index 4b392af400ad317d24061a29f44dfe15c3543044..db84f37aa4fc3477b17599a48a4de9b45cfb6c1f 100644
--- a/python/paddle/v2/data_set/mnist.py
+++ b/python/paddle/v2/dataset/mnist.py
@@ -3,7 +3,7 @@ import sklearn.model_selection
 import numpy
 from config import DATA_HOME
 
-__all__ = ['MNIST', 'train_creator', 'test_creator']
+__all__ = ['train_creator', 'test_creator']
 
 
 def __mnist_reader_creator__(data, target):