diff --git a/python/paddle/distributed/fleet/launch_utils.py b/python/paddle/distributed/fleet/launch_utils.py
index ee5eb807fad7019cd2b7dff85170325a6d8230d1..4b1eef72ee917782e2073af9cbd1b4678169bd69 100644
--- a/python/paddle/distributed/fleet/launch_utils.py
+++ b/python/paddle/distributed/fleet/launch_utils.py
@@ -29,6 +29,7 @@ import struct
 
 import paddle
 import paddle.fluid as fluid
+from distutils.util import strtobool
 logger = logging.getLogger("root")
 logger.propagate = False
 
@@ -349,7 +350,7 @@ def add_arguments(argname, type, default, help, argparser, **kwargs):
         add_argument("name", str, "Jonh", "User name.", parser)
         args = parser.parse_args()
     """
-    type = distutils.util.strtobool if type == bool else type
+    type = strtobool if type == bool else type
     argparser.add_argument(
         "--" + argname,
         default=default,
@@ -685,7 +686,7 @@ def get_device_proc_info(args):
         gpus = get_gpus(args.gpus)
         if args.nproc_per_node is not None:
             assert (len(gpus) % int(args.nproc_per_node)) ==0, \
-                "gpus' number:{} mod args.nproc_per_node:{} must == 0".format(len(gpus), arg.nproc_per_node)
+                "gpus' number:{} mod args.nproc_per_node:{} must == 0".format(len(gpus), args.nproc_per_node)
 
             n = int(len(gpus) / int(args.nproc_per_node))
             devices_per_proc = [
@@ -699,7 +700,7 @@ def get_device_proc_info(args):
         xpus = get_xpus(args.xpus)
         if args.nproc_per_node is not None:
             assert (len(xpus) % int(args.nproc_per_node)) == 0, \
-                "xpus' number:{} mod args.nproc_per_node:{} must == 0".format(len(xpus), arg.nproc_per_node)
+                "xpus' number:{} mod args.nproc_per_node:{} must == 0".format(len(xpus), args.nproc_per_node)
 
             n = int(len(xpus) / int(args.nproc_per_node))
             devices_per_proc = [
diff --git a/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_1d.py b/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_1d.py
index ee4b2c002f5afaf390b42e13e5cf7f34906cd90a..6ebc89b18738c78fcf2a407910de38511b7afa1e 100644
--- a/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_1d.py
+++ b/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_1d.py
@@ -15,6 +15,7 @@
 
 from __future__ import print_function
 
+import unittest
 import paddle
 from paddle.fluid.contrib import sparsity
 from paddle.fluid.tests.unittests.asp.asp_pruning_base import TestASPHelperPruningBase
diff --git a/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_2d_greedy.py b/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_2d_greedy.py
index 4bdd310f0209a94f639b107e7279726e196e6a7d..8ec8ab485250e065993a01236a14cdc4eb4261cf 100644
--- a/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_2d_greedy.py
+++ b/python/paddle/fluid/tests/unittests/asp/test_asp_pruning_2d_greedy.py
@@ -15,6 +15,7 @@
 
 from __future__ import print_function
 
+import unittest
 import paddle
 from paddle.fluid.contrib import sparsity
 from paddle.fluid.tests.unittests.asp.asp_pruning_base import TestASPHelperPruningBase
diff --git a/python/paddle/fluid/tests/unittests/parallel_dygraph_transformer.py b/python/paddle/fluid/tests/unittests/parallel_dygraph_transformer.py
index cfc2ccd4cf7beecd0867582f5cdb5ca8da7a43c5..f149637641add4ef9178aacbc4d6b265b663ca72 100644
--- a/python/paddle/fluid/tests/unittests/parallel_dygraph_transformer.py
+++ b/python/paddle/fluid/tests/unittests/parallel_dygraph_transformer.py
@@ -20,6 +20,7 @@ import six
 import paddle
 import paddle.fluid as fluid
 from paddle.fluid.dygraph import Embedding, LayerNorm, Linear, to_variable, Layer
+from paddle.optimizer.lr import NoamDecay
 
 from test_dist_base import runtime_main, TestParallelDyGraphRunnerBase
 """
diff --git a/python/paddle/fluid/tests/unittests/xpu/test_batch_norm_op_xpu.py b/python/paddle/fluid/tests/unittests/xpu/test_batch_norm_op_xpu.py
index 1cdec863b2ac3d0b4d845e1f59aebe57baa6d4c2..8132a78f696756eebf09b4dee4fe8e47aed34f88 100644
--- a/python/paddle/fluid/tests/unittests/xpu/test_batch_norm_op_xpu.py
+++ b/python/paddle/fluid/tests/unittests/xpu/test_batch_norm_op_xpu.py
@@ -141,7 +141,7 @@ class TestXPUBatchNormOp(unittest.TestCase):
         else:
             raise ValueError(
                 "Unsupported data layout! Only NCHW and NHWC is supported, but received "
-                + data_layout)
+                + self.data_layout)
         np.random.seed(1024)
         self.x_np = np.random.random_sample(self.shape).astype(self.dtype)
         self.scale_np = np.random.random_sample(
diff --git a/tools/CrossStackProfiler/NetFileReader.py b/tools/CrossStackProfiler/NetFileReader.py
index 29c2ae85e60458f6b712e3eeadb739a2dee70d09..fe900fab2ad24bd6b52ec71d0e0d44fe38adf28b 100755
--- a/tools/CrossStackProfiler/NetFileReader.py
+++ b/tools/CrossStackProfiler/NetFileReader.py
@@ -17,6 +17,7 @@ import json
 import glob
 import logging
 import pandas as pd
+import multiprocessing
 
 from multiprocessing import Process