diff --git a/paddlespeech/s2t/decoders/__init__.py b/paddlespeech/s2t/decoders/__init__.py
index f04a6d19..185a92b8 100644
--- a/paddlespeech/s2t/decoders/__init__.py
+++ b/paddlespeech/s2t/decoders/__init__.py
@@ -11,4 +11,3 @@
 # 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 .ctcdecoder import swig_wrapper
diff --git a/paddlespeech/s2t/models/ds2/__init__.py b/paddlespeech/s2t/models/ds2/__init__.py
index 39bea5bf..efa50863 100644
--- a/paddlespeech/s2t/models/ds2/__init__.py
+++ b/paddlespeech/s2t/models/ds2/__init__.py
@@ -13,5 +13,17 @@
 # limitations under the License.
 from .deepspeech2 import DeepSpeech2InferModel
 from .deepspeech2 import DeepSpeech2Model
+from paddlespeech.s2t.utils import dynamic_pip_install
+
+try:
+    import swig_decoders
+except:
+    try:
+        package_name = 'paddlespeech_ctcdecoders'
+        dynamic_pip_install.install(package_name)
+    except:
+        raise RuntimeError(
+            "Can not install package paddlespeech_ctcdecoders on your system. \
+                The DeepSpeech2 model is not supported for your system")
 
 __all__ = ['DeepSpeech2Model', 'DeepSpeech2InferModel']
diff --git a/paddlespeech/s2t/models/ds2_online/__init__.py b/paddlespeech/s2t/models/ds2_online/__init__.py
index 255000ee..65ddd512 100644
--- a/paddlespeech/s2t/models/ds2_online/__init__.py
+++ b/paddlespeech/s2t/models/ds2_online/__init__.py
@@ -13,5 +13,17 @@
 # limitations under the License.
 from .deepspeech2 import DeepSpeech2InferModelOnline
 from .deepspeech2 import DeepSpeech2ModelOnline
+from paddlespeech.s2t.utils import dynamic_pip_install
+
+try:
+    import swig_decoders
+except:
+    try:
+        package_name = 'paddlespeech_ctcdecoders'
+        dynamic_pip_install.install(package_name)
+    except:
+        raise RuntimeError(
+            "Can not install package paddlespeech_ctcdecoders on your system. \
+                The DeepSpeech2 model is not supported for your system")
 
 __all__ = ['DeepSpeech2ModelOnline', 'DeepSpeech2InferModelOnline']
diff --git a/paddlespeech/s2t/modules/ctc.py b/paddlespeech/s2t/modules/ctc.py
index 120abd2b..774bcc62 100644
--- a/paddlespeech/s2t/modules/ctc.py
+++ b/paddlespeech/s2t/modules/ctc.py
@@ -28,8 +28,20 @@ try:
     from paddlespeech.s2t.decoders.ctcdecoder.swig_wrapper import ctc_beam_search_decoder_batch  # noqa: F401
     from paddlespeech.s2t.decoders.ctcdecoder.swig_wrapper import ctc_greedy_decoder  # noqa: F401
     from paddlespeech.s2t.decoders.ctcdecoder.swig_wrapper import Scorer  # noqa: F401
-except Exception as e:
-    logger.info("ctcdecoder not installed!")
+except:
+    try:
+        from paddlespeech.s2t.utils import dynamic_pip_install
+        package_name = 'paddlespeech_ctcdecoders'
+        dynamic_pip_install.install(package_name)
+        from paddlespeech.s2t.decoders.ctcdecoder.swig_wrapper import ctc_beam_search_decoder_batch  # noqa: F401
+        from paddlespeech.s2t.decoders.ctcdecoder.swig_wrapper import ctc_greedy_decoder  # noqa: F401
+        from paddlespeech.s2t.decoders.ctcdecoder.swig_wrapper import Scorer  # noqa: F401
+    except Exception as e:
+        logger.info("paddlespeech_ctcdecoders not installed!")
+
+#try:
+#except Exception as e:
+#    logger.info("ctcdecoder not installed!")
 
 __all__ = ['CTCDecoder']
 
@@ -51,7 +63,7 @@ class CTCDecoderBase(nn.Layer):
             dropout_rate (float): dropout rate (0.0 ~ 1.0)
             reduction (bool): reduce the CTC loss into a scalar, True for 'sum' or 'none'
             batch_average (bool): do batch dim wise average.
-            grad_norm_type (str): Default, None. one of 'instance', 'batch', 'frame', None. 
+            grad_norm_type (str): Default, None. one of 'instance', 'batch', 'frame', None.
         """
         assert check_argument_types()
         super().__init__()
diff --git a/paddlespeech/s2t/utils/dynamic_pip_install.py b/paddlespeech/s2t/utils/dynamic_pip_install.py
new file mode 100644
index 00000000..39e9c35f
--- /dev/null
+++ b/paddlespeech/s2t/utils/dynamic_pip_install.py
@@ -0,0 +1,22 @@
+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# 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.
+import pip
+
+
+def install(package_name):
+    if int(pip.__version__.split('.')[0]) > 9:
+        from pip._internal import main
+    else:
+        from pip import main
+    main(['install', package_name])
diff --git a/setup.py b/setup.py
index c1ae00ce..3841183c 100644
--- a/setup.py
+++ b/setup.py
@@ -45,7 +45,6 @@ requirements = {
         "nltk",
         "pandas",
         "paddleaudio",
-        "paddlespeech_ctcdecoders",
         "paddlespeech_feat",
         "praatio~=4.1",
         "pypi-kenlm",
@@ -71,6 +70,7 @@ requirements = {
         "ConfigArgParse",
         "coverage",
         "gpustat",
+        "paddlespeech_ctcdecoders",
         "phkit",
         "Pillow",
         "pybind11",