diff --git a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/README.md b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/README.md
index 08533849bf00a8dbb11db70af7cf051db0ebf6ed..b7dba8ceda67875bfce020c2dac52e6c7bdb0d8b 100644
--- a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/README.md
+++ b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/README.md
@@ -19,6 +19,16 @@ $ hub run chinese_ocr_db_crnn_mobile --input_path "/PATH/TO/IMAGE"
 
 ## API
 
+### \_\_init\_\_(text_detector_module=None, enable_mkldnn=False)
+
+构造ChineseOCRDBCRNN对象
+
+**参数**
+
+* text_detector_module(str): 文字检测PaddleHub Module名字，如设置为None，则默认使用[chinese_text_detection_db_mobile Module](https://www.paddlepaddle.org.cn/hubdetail?name=chinese_text_detection_db_mobile&en_category=TextRecognition)。其作用为检测图片当中的文本。
+* enable_mkldnn(bool): 是否开启mkldnn加速CPU计算。该参数仅在CPU运行下设置有效。默认为False。
+
+
 ```python
 def recognize_text(images=[],
                     paths=[],
diff --git a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/module.py b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/module.py
index 490bba80765e5c15099001ce5abe65f8e504ab1a..a3e4bd5bd366ccc89666ed8b6acb63b830359202 100644
--- a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/module.py
+++ b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_mobile/module.py
@@ -25,14 +25,14 @@ from chinese_ocr_db_crnn_mobile.utils import base64_to_cv2, draw_ocr, get_image_
 
 @moduleinfo(
     name="chinese_ocr_db_crnn_mobile",
-    version="1.0.3",
+    version="1.0.4",
     summary=
     "The module can recognize the chinese texts in an image. Firstly, it will detect the text box positions based on the differentiable_binarization_chn module. Then it recognizes the chinese texts. ",
     author="paddle-dev",
     author_email="paddle-dev@baidu.com",
     type="cv/text_recognition")
 class ChineseOCRDBCRNN(hub.Module):
-    def _initialize(self, text_detector_module=None):
+    def _initialize(self, text_detector_module=None, enable_mkldnn=False):
         """
         initialize with the necessary elements
         """
@@ -49,6 +49,7 @@ class ChineseOCRDBCRNN(hub.Module):
         self.font_file = os.path.join(self.directory, 'assets', 'simfang.ttf')
         self.pretrained_model_path = os.path.join(self.directory,
                                                   'inference_model')
+        self.enable_mkldnn = enable_mkldnn
         self._set_config()
 
     def _set_config(self):
@@ -70,6 +71,8 @@ class ChineseOCRDBCRNN(hub.Module):
             config.enable_use_gpu(8000, 0)
         else:
             config.disable_gpu()
+            if self.enable_mkldnn:
+                config.enable_mkldnn()
 
         config.disable_glog_info()
 
@@ -92,7 +95,9 @@ class ChineseOCRDBCRNN(hub.Module):
         """
         if not self._text_detector_module:
             self._text_detector_module = hub.Module(
-                name='chinese_text_detection_db_mobile')
+                name='chinese_text_detection_db_mobile',
+                enable_mkldnn=self.enable_mkldnn,
+                version='1.0.2')
         return self._text_detector_module
 
     def read_images(self, paths=[]):
diff --git a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/README.md b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/README.md
index 389dbb71ba4cd31fe78ef6abef0505329a7be5c9..2b0aa7d7a4a9ecf26fde0e1e197699c0f6b33e50 100644
--- a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/README.md
+++ b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/README.md
@@ -19,6 +19,15 @@ $ hub run chinese_ocr_db_crnn_server --input_path "/PATH/TO/IMAGE"
 
 ## API
 
+### \_\_init\_\_(text_detector_module=None, enable_mkldnn=False)
+
+构造ChineseOCRDBCRNNServer对象
+
+**参数**
+
+* text_detector_module(str): 文字检测PaddleHub Module名字，如设置为None，则默认使用[chinese_text_detection_db_server Module](https://www.paddlepaddle.org.cn/hubdetail?name=chinese_text_detection_db_server&en_category=TextRecognition)。其作用为检测图片当中的文本。
+* enable_mkldnn(bool): 是否开启mkldnn加速CPU计算。该参数仅在CPU运行下设置有效。默认为False。
+
 ```python
 def recognize_text(images=[],
                     paths=[],
diff --git a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/module.py b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/module.py
index 2c82cdba4976209b1726a17df8687ed875e9b830..88422344bbd05f1e88d64c3af73f6246c69806e0 100644
--- a/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/module.py
+++ b/hub_module/modules/image/text_recognition/chinese_ocr_db_crnn_server/module.py
@@ -25,14 +25,14 @@ from chinese_ocr_db_crnn_server.utils import base64_to_cv2, draw_ocr, get_image_
 
 @moduleinfo(
     name="chinese_ocr_db_crnn_server",
-    version="1.0.2",
+    version="1.0.3",
     summary=
     "The module can recognize the chinese texts in an image. Firstly, it will detect the text box positions based on the differentiable_binarization_chn module. Then it recognizes the chinese texts. ",
     author="paddle-dev",
     author_email="paddle-dev@baidu.com",
     type="cv/text_recognition")
 class ChineseOCRDBCRNNServer(hub.Module):
-    def _initialize(self, text_detector_module=None):
+    def _initialize(self, text_detector_module=None, enable_mkldnn=False):
         """
         initialize with the necessary elements
         """
@@ -49,6 +49,8 @@ class ChineseOCRDBCRNNServer(hub.Module):
         self.font_file = os.path.join(self.directory, 'assets', 'simfang.ttf')
         self.pretrained_model_path = os.path.join(self.directory, 'assets',
                                                   'ch_rec_r34_vd_crnn')
+        self.enable_mkldnn = enable_mkldnn
+
         self._set_config()
 
     def _set_config(self):
@@ -70,6 +72,8 @@ class ChineseOCRDBCRNNServer(hub.Module):
             config.enable_use_gpu(8000, 0)
         else:
             config.disable_gpu()
+            if self.enable_mkldnn:
+                config.enable_mkldnn()
 
         config.disable_glog_info()
 
@@ -92,7 +96,9 @@ class ChineseOCRDBCRNNServer(hub.Module):
         """
         if not self._text_detector_module:
             self._text_detector_module = hub.Module(
-                name='chinese_text_detection_db_server')
+                name='chinese_text_detection_db_server',
+                enable_mkldnn=self.enable_mkldnn,
+                version='1.0.1')
         return self._text_detector_module
 
     def read_images(self, paths=[]):
@@ -423,7 +429,7 @@ class ChineseOCRDBCRNNServer(hub.Module):
 
 
 if __name__ == '__main__':
-    ocr = ChineseOCRDBCRNNServer()
+    ocr = ChineseOCRDBCRNNServer(enable_mkldnn=True)
     print(ocr.name)
     image_path = [
         '/mnt/zhangxuefei/PaddleOCR/doc/imgs/11.jpg',
diff --git a/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/README.md b/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/README.md
index 8dd066c5c1733a77e1257a1e8190ffa31209dc35..6c43493ff553bd1366f35530e916170a9841cab2 100644
--- a/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/README.md
+++ b/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/README.md
@@ -19,6 +19,17 @@ $ hub run chinese_text_detection_db_mobile --input_path "/PATH/TO/IMAGE"
 
 ## API
 
+## API
+
+### \_\_init\_\_(enable_mkldnn=False)
+
+构造ChineseTextDetectionDB对象
+
+**参数**
+
+* enable_mkldnn(bool): 是否开启mkldnn加速CPU计算。该参数仅在CPU运行下设置有效。默认为False。
+
+
 ```python
 def detect_text(paths=[],
                 images=[],
@@ -51,7 +62,7 @@ def detect_text(paths=[],
 import paddlehub as hub
 import cv2
 
-text_detector = hub.Module(name="chinese_text_detection_db_mobile")
+text_detector = hub.Module(name="chinese_text_detection_db_mobile", enable_mk)
 result = text_detector.detect_text(images=[cv2.imread('/PATH/TO/IMAGE')])
 
 # or
diff --git a/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/module.py b/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/module.py
index 0ee5756f5bece95c08ecca2cd9bcdfe8f5ac53b0..be00f7e2a0621df619f96f9e671a2061f171b925 100644
--- a/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/module.py
+++ b/hub_module/modules/image/text_recognition/chinese_text_detection_db_mobile/module.py
@@ -29,19 +29,21 @@ def base64_to_cv2(b64str):
 
 @moduleinfo(
     name="chinese_text_detection_db_mobile",
-    version="1.0.1",
+    version="1.0.2",
     summary=
     "The module aims to detect chinese text position in the image, which is based on differentiable_binarization algorithm.",
     author="paddle-dev",
     author_email="paddle-dev@baidu.com",
     type="cv/text_recognition")
 class ChineseTextDetectionDB(hub.Module):
-    def _initialize(self):
+    def _initialize(self, enable_mkldnn=False):
         """
         initialize with the necessary elements
         """
         self.pretrained_model_path = os.path.join(self.directory,
                                                   'inference_model')
+        self.enable_mkldnn = enable_mkldnn
+
         self._set_config()
 
     def check_requirements(self):
@@ -71,6 +73,8 @@ class ChineseTextDetectionDB(hub.Module):
             config.enable_use_gpu(8000, 0)
         else:
             config.disable_gpu()
+            if self.enable_mkldnn:
+                config.enable_mkldnn()
 
         config.disable_glog_info()
 
diff --git a/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/README.md b/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/README.md
index 37a70e19de5e0fdba4a9a1f1b185748e0ff629a5..393eec37534ed75c49f09fd24394617b8f728f23 100644
--- a/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/README.md
+++ b/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/README.md
@@ -19,6 +19,14 @@ $ hub run chinese_text_detection_db_server --input_path "/PATH/TO/IMAGE"
 
 ## API
 
+### \_\_init\_\_(enable_mkldnn=False)
+
+构造ChineseTextDetectionDBServer对象
+
+**参数**
+
+* enable_mkldnn(bool): 是否开启mkldnn加速CPU计算。该参数仅在CPU运行下设置有效。默认为False。
+
 ```python
 def detect_text(paths=[],
                 images=[],
diff --git a/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/module.py b/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/module.py
index 6c07144b790c4b4e136e40147a201cc2ee926b82..7baec43d9acba27c711546360e3e280f07cce4e4 100644
--- a/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/module.py
+++ b/hub_module/modules/image/text_recognition/chinese_text_detection_db_server/module.py
@@ -29,19 +29,21 @@ def base64_to_cv2(b64str):
 
 @moduleinfo(
     name="chinese_text_detection_db_server",
-    version="1.0.0",
+    version="1.0.1",
     summary=
     "The module aims to detect chinese text position in the image, which is based on differentiable_binarization algorithm.",
     author="paddle-dev",
     author_email="paddle-dev@baidu.com",
     type="cv/text_recognition")
 class ChineseTextDetectionDBServer(hub.Module):
-    def _initialize(self):
+    def _initialize(self, enable_mkldnn=False):
         """
         initialize with the necessary elements
         """
         self.pretrained_model_path = os.path.join(self.directory,
                                                   'ch_det_r50_vd_db')
+        self.enable_mkldnn = enable_mkldnn
+
         self._set_config()
 
     def check_requirements(self):
@@ -71,6 +73,8 @@ class ChineseTextDetectionDBServer(hub.Module):
             config.enable_use_gpu(8000, 0)
         else:
             config.disable_gpu()
+            if self.enable_mkldnn:
+                config.enable_mkldnn()
 
         config.disable_glog_info()