diff --git a/python/paddle_serving_app/__init__.py b/python/paddle_serving_app/__init__.py
index 2a6225570c3de61ba6e0a0587f81175816cd0f8d..11ad09a1d880a8b235e5cf1b99f6be91ec9cccbf 100644
--- a/python/paddle_serving_app/__init__.py
+++ b/python/paddle_serving_app/__init__.py
@@ -11,10 +11,4 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from .reader.chinese_bert_reader import ChineseBertReader
-from .reader.image_reader import ImageReader, File2Image, URL2Image, Sequential, Normalize, CenterCrop, Resize, PadStride
-from .reader.lac_reader import LACReader
-from .reader.senta_reader import SentaReader
-from .reader.imdb_reader import IMDBDataset
 from .models import ServingModels
-from .local_predict import Debugger
diff --git a/python/paddle_serving_app/reader/lac_reader.py b/python/paddle_serving_app/reader/lac_reader.py
index 720bbf9c61051dcdc877f0a1f4933718be32263d..7e804ff371e2d90d79f7f663e83a854b1b0c9647 100644
--- a/python/paddle_serving_app/reader/lac_reader.py
+++ b/python/paddle_serving_app/reader/lac_reader.py
@@ -48,10 +48,16 @@ def load_kv_dict(dict_path,
 class LACReader(object):
     """data reader"""
 
-    def __init__(self, dict_folder):
+    def __init__(self, dict_folder=""):
         # read dict
         #basepath = os.path.abspath(__file__)
         #folder = os.path.dirname(basepath)
+        if dict_folder == "":
+            dict_folder = "lac_dict"
+            if not os.path.exists(dict_folder):
+                r = os.system(
+                    "wget https://paddle-serving.bj.bcebos.com/reader/lac/lac_dict.tar.gz  --no-check-certificate && tar -xzvf lac_dict.tar.gz"
+                )
         word_dict_path = os.path.join(dict_folder, "word.dic")
         label_dict_path = os.path.join(dict_folder, "tag.dic")
         replace_dict_path = os.path.join(dict_folder, "q2b.dic")
diff --git a/python/paddle_serving_app/reader/senta_reader.py b/python/paddle_serving_app/reader/senta_reader.py
index 6e608b822fbb66f11288ea0080c8e264d8e5c34a..e0c93c00d1a6acb0c3d30294d40fb63b4929a639 100644
--- a/python/paddle_serving_app/reader/senta_reader.py
+++ b/python/paddle_serving_app/reader/senta_reader.py
@@ -14,10 +14,11 @@
 
 import sys
 import io
+import os
 
 
 class SentaReader():
-    def __init__(self, vocab_path, max_seq_len=20):
+    def __init__(self, vocab_path="", max_seq_len=20):
         self.max_seq_len = max_seq_len
         self.word_dict = self.load_vocab(vocab_path)
 
@@ -25,6 +26,13 @@ class SentaReader():
         """
         load the given vocabulary
         """
+        if vocab_path == "":
+            vocab_path = "senta_vocab.txt"
+            if not os.path.exists(vocab_path):
+                r = os.system(
+                    " wget https://paddle-serving.bj.bcebos.com/reader/senta/senta_vocab.txt --no-check-certificate"
+                )
+
         vocab = {}
         with io.open(vocab_path, 'r', encoding='utf8') as f:
             for line in f: