diff --git a/fluid/DeepASR/data_utils/data_reader.py b/fluid/DeepASR/data_utils/data_reader.py
index b08e7f55a45d7e0dd1550ad84c388c49867c93c1..f7bc9c6602f82c05fb7eff9a9c90b43f88cd02b1 100644
--- a/fluid/DeepASR/data_utils/data_reader.py
+++ b/fluid/DeepASR/data_utils/data_reader.py
@@ -225,8 +225,8 @@ class DataReader(object):
         @suppress_complaints(verbose=self._verbose)
         def ordered_processing_task(sample_info_queue, sample_queue, out_order):
             if self._verbose == 0:
-                signal.signal(signal.SIGTERM, suppress_signal())
-                signal.signal(signal.SIGINT, suppress_signal())
+                signal.signal(signal.SIGTERM, suppress_signal)
+                signal.signal(signal.SIGINT, suppress_signal)
 
             def read_bytes(fpath, start, size):
                 f = open(fpath, 'r')
diff --git a/fluid/DeepASR/model.py b/fluid/DeepASR/model_utils/model.py
similarity index 92%
rename from fluid/DeepASR/model.py
rename to fluid/DeepASR/model_utils/model.py
index 226a06b0a9673d38893341766c94deb36bf9f64b..4e29394a0a2ced700c99f3ea9fa516a9a58e1420 100644
--- a/fluid/DeepASR/model.py
+++ b/fluid/DeepASR/model_utils/model.py
@@ -9,12 +9,16 @@ import paddle.v2.fluid as fluid
 def stacked_lstmp_model(hidden_dim,
                         proj_dim,
                         stacked_num,
+                        class_num,
                         parallel=False,
-                        is_train=True,
-                        class_num=1749):
+                        is_train=True):
     """ The model for DeepASR. The main structure is composed of stacked 
         identical LSTMP (LSTM with recurrent projection) layers.
 
+        When running in training and validation phase, the feeding dictionary
+        is {'feature', 'label'}, fed by the LodTensor for feature data and 
+        label data respectively. And in inference, only `feature` is needed.
+
     Args:
 	hidden_dim(int): The hidden state's dimension of the LSTMP layer.
 	proj_dim(int): The projection size of the LSTMP layer.
diff --git a/fluid/DeepASR/tools/_init_paths.py b/fluid/DeepASR/tools/_init_paths.py
new file mode 100644
index 0000000000000000000000000000000000000000..228dbae6bf95231030c1858c4d30b49f162f46e2
--- /dev/null
+++ b/fluid/DeepASR/tools/_init_paths.py
@@ -0,0 +1,19 @@
+"""Add the parent directory to $PYTHONPATH"""
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import os.path
+import sys
+
+
+def add_path(path):
+    if path not in sys.path:
+        sys.path.insert(0, path)
+
+
+this_dir = os.path.dirname(__file__)
+
+# Add project path to PYTHONPATH
+proj_path = os.path.join(this_dir, '..')
+add_path(proj_path)
diff --git a/fluid/DeepASR/profile.py b/fluid/DeepASR/tools/profile.py
similarity index 95%
rename from fluid/DeepASR/profile.py
rename to fluid/DeepASR/tools/profile.py
index 9f9cce0411159e5e2f20f0eb091a07d5dfe42caf..a45ce98acb3b70bfa8292c05fd52c85a3a3ddc48 100644
--- a/fluid/DeepASR/profile.py
+++ b/fluid/DeepASR/tools/profile.py
@@ -7,14 +7,14 @@ import numpy as np
 import argparse
 import time
 
-import paddle.v2 as paddle
 import paddle.v2.fluid as fluid
 import paddle.v2.fluid.profiler as profiler
+import _init_paths
 import data_utils.augmentor.trans_mean_variance_norm as trans_mean_variance_norm
 import data_utils.augmentor.trans_add_delta as trans_add_delta
 import data_utils.augmentor.trans_splice as trans_splice
 import data_utils.data_reader as reader
-from model import stacked_lstmp_model
+from model_utils.model import stacked_lstmp_model
 from data_utils.util import lodtensor_to_ndarray
 
 
@@ -118,8 +118,12 @@ def profile(args):
         raise ValueError(
             "arg 'first_batches_to_skip' must not be smaller than 0.")
 
-    _, avg_cost, accuracy = stacked_lstmp_model(args.hidden_dim, args.proj_dim,
-                                                args.stacked_num, args.parallel)
+    _, avg_cost, accuracy = stacked_lstmp_model(
+        hidden_dim=args.hidden_dim,
+        proj_dim=args.proj_dim,
+        stacked_num=args.stacked_num,
+        class_num=1749,
+        parallel=args.parallel)
 
     adam_optimizer = fluid.optimizer.Adam(learning_rate=args.learning_rate)
     adam_optimizer.minimize(avg_cost)
@@ -173,7 +177,6 @@ def profile(args):
             else:
                 sys.stdout.write('.')
                 sys.stdout.flush()
-
         time_consumed = time.time() - start_time
         frames_per_sec = frames_seen / time_consumed
         print("\nTime consumed: %f s, performance: %f frames/s." %
diff --git a/fluid/DeepASR/train.py b/fluid/DeepASR/train.py
index b2584259742a4a995c2bb4dadf3b0b9acc80ec15..1c45f0a086332288760c359d9845aa94641cf7d7 100644
--- a/fluid/DeepASR/train.py
+++ b/fluid/DeepASR/train.py
@@ -8,15 +8,13 @@ import numpy as np
 import argparse
 import time
 
-import paddle.v2 as paddle
 import paddle.v2.fluid as fluid
-import paddle.v2.fluid.profiler as profiler
 import data_utils.augmentor.trans_mean_variance_norm as trans_mean_variance_norm
 import data_utils.augmentor.trans_add_delta as trans_add_delta
 import data_utils.augmentor.trans_splice as trans_splice
 import data_utils.data_reader as reader
 from data_utils.util import lodtensor_to_ndarray
-from model import stacked_lstmp_model
+from model_utils.model import stacked_lstmp_model
 
 
 def parse_args():
@@ -116,8 +114,14 @@ def train(args):
     """train in loop.
     """
 
+    # prediction, avg_cost, accuracy = stacked_lstmp_model(args.hidden_dim, 
+    #    args.proj_dim, args.stacked_num, class_num=1749, args.parallel)
     prediction, avg_cost, accuracy = stacked_lstmp_model(
-        args.hidden_dim, args.proj_dim, args.stacked_num, args.parallel)
+        hidden_dim=args.hidden_dim,
+        proj_dim=args.proj_dim,
+        stacked_num=args.stacked_num,
+        class_num=1749,
+        parallel=args.parallel)
 
     adam_optimizer = fluid.optimizer.Adam(learning_rate=args.learning_rate)
     adam_optimizer.minimize(avg_cost)