diff --git a/python/paddle/tests/test_datasets.py b/python/paddle/tests/test_datasets.py
index d119d2c5ccad62f85929327e4783b08b7108b63f..3aa21ae2db267d40d4483b27ed89b0da349ce0db 100644
--- a/python/paddle/tests/test_datasets.py
+++ b/python/paddle/tests/test_datasets.py
@@ -179,6 +179,13 @@ class TestFASHIONMNISTTrain(unittest.TestCase):
         with self.assertRaises(ValueError):
             mnist = FashionMNIST(mode='train', transform=transform, backend=1)
 
+    def test_dataset_value(self):
+        fmnist = FashionMNIST(mode='train')
+        value = np.mean([np.array(x[0]) for x in fmnist])
+
+        # 72.94035223214286 was getted from competitive products
+        np.testing.assert_allclose(value, 72.94035223214286)
+
 
 class TestFlowersTrain(unittest.TestCase):
     def test_main(self):
diff --git a/python/paddle/vision/datasets/mnist.py b/python/paddle/vision/datasets/mnist.py
index 3d752ece346b7a477c385f2a06f96d696b4b9eb0..0f4d4947aa5f8b5a9cdcbac8642e0ac73e14a905 100644
--- a/python/paddle/vision/datasets/mnist.py
+++ b/python/paddle/vision/datasets/mnist.py
@@ -163,7 +163,7 @@ class MNIST(Dataset):
         image = np.reshape(image, [28, 28])
 
         if self.backend == 'pil':
-            image = Image.fromarray(image, mode='L')
+            image = Image.fromarray(image.astype('uint8'), mode='L')
 
         if self.transform is not None:
             image = self.transform(image)