[CodeStyle][isort] introducing `isort` (part1) (#46475)

* add isort config

* isort all files

.pre-commit-config.yaml

@@ -60,6 +60,10 @@ repos:
     hooks:
     -   id: black
         files: (.*\.(py|pyi|bzl)|BUILD|.*\.BUILD|WORKSPACE)$
+-   repo: https://github.com/pycqa/isort
+    rev: 5.10.1
+    hooks:
+    -   id: isort
 -   repo: https://github.com/PyCQA/flake8
     rev: 4.0.1
     hooks:

pyproject.toml

@@ -2,3 +2,30 @@
 exclude = "build"
 line-length = 80
 skip-string-normalization = true
+
+[tool.isort]
+profile = "black"
+line_length = 80
+known_first_party = ["paddle"]
+skip = ["build", "__init__.py"]
+extend_skip_glob = [
+    # These files do not need to be formatted,
+    # see .flake8 for more details
+    "python/paddle/fluid/[!t]**",
+    "python/paddle/fluid/tra**",
+    "*_pb2.py",
+    "python/paddle/utils/gast/**",
+    "python/paddle/fluid/tests/unittests/npu/**",
+    "python/paddle/fluid/tests/unittests/mlu/**",
+
+    # These files will be fixed in the future
+    "cmake/**",
+    "paddle/**",
+    "r/**",
+    "tools/**",
+    "python/paddle/[!f]**",
+    "python/paddle/fluid/tests/unittests/[t-z]**",
+    "python/paddle/fluid/tests/unittests/dygraph_to_static/test_error.py",
+    "python/paddle/fluid/tests/unittests/dygraph_to_static/**",
+    "python/paddle/fluid/tests/unittests/ipu/test_dy2static_ipu.py",
+]

python/paddle/fft.py

@@ -13,14 +13,17 @@
 # limitations under the License.
 
 from typing import Sequence
+
 import numpy as np
+
 import paddle
-from .tensor.attribute import is_floating_point, is_integer
-from .tensor.creation import _real_to_complex_dtype, _complex_to_real_dtype
-from .fluid.framework import _in_legacy_dygraph, in_dygraph_mode
+
 from . import _C_ops, _legacy_C_ops
 from .fluid.data_feeder import check_variable_and_dtype
+from .fluid.framework import _in_legacy_dygraph, in_dygraph_mode
 from .fluid.layer_helper import LayerHelper
+from .tensor.attribute import is_floating_point, is_integer
+from .tensor.creation import _complex_to_real_dtype, _real_to_complex_dtype
 
 __all__ = [
     'fft',

python/paddle/fluid/tests/book/notest_understand_sentiment.py

@@ -12,14 +12,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import paddle.fluid as fluid
-import paddle
 import contextlib
 import math
-import numpy as np
-import sys
 import os
+import sys
+import unittest
+
+import numpy as np
+
+import paddle
+import paddle.fluid as fluid
 
 
 def convolution_net(

python/paddle/fluid/tests/book/test_fit_a_line.py

@@ -12,18 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle
-import paddle.fluid as fluid
-import paddle.static.amp as amp
-
 import contextlib
-import numpy
-import unittest
 import math
-import sys
 import os
 import struct
+import sys
 import tempfile
+import unittest
+
+import numpy
+
+import paddle
+import paddle.fluid as fluid
+import paddle.static.amp as amp
 
 paddle.enable_static()
 

python/paddle/fluid/tests/book/test_image_classification.py

@@ -12,15 +12,17 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle
-import paddle.fluid as fluid
 import contextlib
 import math
-import sys
-import numpy
-import unittest
 import os
+import sys
 import tempfile
+import unittest
+
+import numpy
+
+import paddle
+import paddle.fluid as fluid
 
 paddle.enable_static()
 

python/paddle/fluid/tests/book/test_label_semantic_roles.py

@@ -13,11 +13,12 @@
 # limitations under the License.
 
 import contextlib
-import numpy as np
 import os
+import tempfile
 import time
 import unittest
-import tempfile
+
+import numpy as np
 
 import paddle
 import paddle.dataset.conll05 as conll05

python/paddle/fluid/tests/book/test_machine_translation.py

@@ -13,15 +13,16 @@
 # limitations under the License.
 
 import contextlib
+import os
+import unittest
 
 import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.framework as framework
 import paddle.fluid.layers as pd
 from paddle.fluid.executor import Executor
-import unittest
-import os
 
 paddle.enable_static()
 

python/paddle/fluid/tests/book/test_recognize_digits.py

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.fluid.core as core
 import math
 import os
 import sys
@@ -22,6 +21,7 @@ import numpy
 
 import paddle
 import paddle.fluid as fluid
+import paddle.fluid.core as core
 
 paddle.enable_static()
 

python/paddle/fluid/tests/book/test_recommender_system.py

@@ -13,15 +13,17 @@
 # limitations under the License.
 
 import math
-import sys
 import os
+import sys
+import tempfile
+
 import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.framework as framework
 import paddle.fluid.layers as layers
 import paddle.fluid.nets as nets
-import tempfile
 from paddle.fluid.executor import Executor
 from paddle.fluid.optimizer import SGDOptimizer
 

python/paddle/fluid/tests/book/test_rnn_encoder_decoder.py

@@ -12,18 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import numpy as np
-import paddle
-import paddle.fluid as fluid
-import paddle.fluid.framework as framework
 import contextlib
 import math
-import sys
 import os
-import unittest
+import sys
 import tempfile
-from paddle.fluid.executor import Executor
+import unittest
+
+import numpy as np
+
 import paddle
+import paddle.fluid as fluid
+import paddle.fluid.framework as framework
+from paddle.fluid.executor import Executor
 
 paddle.enable_static()
 

python/paddle/fluid/tests/book/test_word2vec_book.py

@@ -12,14 +12,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle
-import paddle.fluid as fluid
-import unittest
-import os
-import numpy as np
 import math
+import os
 import sys
 import tempfile
+import unittest
+
+import numpy as np
+
+import paddle
+import paddle.fluid as fluid
 
 paddle.enable_static()
 

python/paddle/fluid/tests/custom_kernel/custom_kernel_dot_c_setup.py

@@ -13,11 +13,13 @@
 # limitations under the License.
 
 import os
-from paddle.fluid import core
+from distutils.core import Extension, setup
 from distutils.sysconfig import get_python_lib
-from distutils.core import setup, Extension
+
 from setuptools.command.build_ext import build_ext
 
+from paddle.fluid import core
+
 
 # refer: https://note.qidong.name/2018/03/setup-warning-strict-prototypes
 # Avoid a gcc warning below:

python/paddle/fluid/tests/custom_kernel/custom_kernel_dot_setup.py

@@ -14,10 +14,12 @@
 
 import os
 import site
-from paddle.fluid import core
-from distutils.core import setup, Extension
+from distutils.core import Extension, setup
+
 from setuptools.command.build_ext import build_ext
 
+from paddle.fluid import core
+
 
 # refer: https://note.qidong.name/2018/03/setup-warning-strict-prototypes
 # Avoid a gcc warning below:

python/paddle/fluid/tests/custom_kernel/test_custom_kernel_dot.py

@@ -15,6 +15,7 @@
 import os
 import sys
 import unittest
+
 import numpy as np
 
 

python/paddle/fluid/tests/custom_kernel/test_custom_kernel_load.py

@@ -13,9 +13,10 @@
 # limitations under the License.
 
 import os
-import sys
 import site
+import sys
 import unittest
+
 import numpy as np
 
 

python/paddle/fluid/tests/custom_op/custom_raw_op_kernel_op_setup.py

@@ -13,10 +13,12 @@
 # limitations under the License.
 
 import os
+
+from utils import extra_compile_args, paddle_includes
+
 import paddle
 import paddle.fluid.core as core
 from paddle.utils.cpp_extension import CppExtension, CUDAExtension, setup
-from utils import paddle_includes, extra_compile_args
 
 if paddle.is_compiled_with_cuda():
     sources = ['custom_raw_op_kernel_op.cc', 'custom_raw_op_kernel_op.cu']

python/paddle/fluid/tests/custom_op/custom_relu_setup.py

@@ -12,8 +12,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from utils import paddle_includes, extra_compile_args, IS_MAC
-from paddle.utils.cpp_extension import CUDAExtension, setup, CppExtension
+from utils import IS_MAC, extra_compile_args, paddle_includes
+
+from paddle.utils.cpp_extension import CppExtension, CUDAExtension, setup
 
 # Mac-CI don't support GPU
 Extension = CppExtension if IS_MAC else CUDAExtension

python/paddle/fluid/tests/custom_op/test_check_abi.py

@@ -12,8 +12,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import os
+import unittest
 import warnings
 
 import paddle.utils.cpp_extension.extension_utils as utils

python/paddle/fluid/tests/custom_op/test_context_pool.py

@@ -14,13 +14,14 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
-from paddle.utils.cpp_extension import load, get_build_directory
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
-from paddle.utils.cpp_extension.extension_utils import run_cmd
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_attrs_jit.py

@@ -14,13 +14,14 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
-from paddle.utils.cpp_extension import load, get_build_directory
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
-from paddle.utils.cpp_extension.extension_utils import run_cmd
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_concat.py

@@ -14,14 +14,15 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
 import paddle.static as static
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_conj.py

@@ -14,14 +14,15 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
 import paddle.static as static
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_linear.py

@@ -14,15 +14,16 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
-import paddle.static as static
 import paddle.nn.functional as F
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
+import paddle.static as static
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_raw_op_kernel_op.py

@@ -12,15 +12,17 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import importlib
 import os
-import paddle
 import shlex
 import site
 import sys
-import importlib
 import unittest
+
 import numpy as np
 
+import paddle
+
 MODULE_NAME = "custom_raw_op_kernel_op_lib"
 
 

python/paddle/fluid/tests/custom_op/test_custom_relu_model.py

@@ -13,18 +13,18 @@
 # limitations under the License.
 
 import os
+import tempfile
 import unittest
+
 import numpy as np
-import tempfile
+from utils import IS_MAC, extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
 from paddle import nn
-from paddle.utils.cpp_extension import load, get_build_directory
+from paddle.fluid.framework import _in_legacy_dygraph, _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
 from paddle.utils.cpp_extension.extension_utils import run_cmd
 
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args, IS_MAC
-from paddle.fluid.framework import _test_eager_guard, _in_legacy_dygraph
-
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.
 file = '{}\\custom_relu_for_model_jit\\custom_relu_for_model_jit.pyd'.format(

python/paddle/fluid/tests/custom_op/test_custom_relu_op_jit.py

@@ -14,13 +14,15 @@
 
 import os
 import unittest
-import paddle
+
 import numpy as np
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import IS_MAC, extra_cc_args, extra_nvcc_args, paddle_includes
 from test_custom_relu_op_setup import custom_relu_dynamic, custom_relu_static
+from utils import IS_MAC, extra_cc_args, extra_nvcc_args, paddle_includes
+
+import paddle
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_relu_op_setup.py

@@ -13,15 +13,17 @@
 # limitations under the License.
 
 import os
-import sys
 import site
+import sys
 import unittest
+
+import numpy as np
+
 import paddle
 import paddle.static as static
-import numpy as np
-from paddle.vision.transforms import Compose, Normalize
-from paddle.utils.cpp_extension.extension_utils import run_cmd
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension.extension_utils import run_cmd
+from paddle.vision.transforms import Compose, Normalize
 
 
 def custom_relu_dynamic(func, device, dtype, np_x, use_func=True):
@@ -318,8 +320,7 @@ class TestNewCustomOpSetUpInstall(unittest.TestCase):
         np_data = np.random.random((1, 1, 28, 28)).astype("float32")
         np_label = np.random.random((1, 1)).astype("int64")
         path_prefix = "custom_op_inference/custom_relu"
-        from paddle.inference import Config
-        from paddle.inference import create_predictor
+        from paddle.inference import Config, create_predictor
 
         for device in self.devices:
             predict = custom_relu_static_inference(

python/paddle/fluid/tests/custom_op/test_custom_simple_slice.py

@@ -14,13 +14,14 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_custom_tanh_double_grad.py

@@ -14,14 +14,15 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, extra_nvcc_args, paddle_includes
 
 import paddle
 import paddle.fluid as fluid
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import paddle_includes, extra_cc_args, extra_nvcc_args
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_dispatch_jit.py

@@ -14,12 +14,14 @@
 
 import os
 import unittest
-import paddle
+
 import numpy as np
-from paddle.utils.cpp_extension import load, get_build_directory
-from utils import paddle_includes, extra_cc_args
-from paddle.utils.cpp_extension.extension_utils import run_cmd
+from utils import extra_cc_args, paddle_includes
+
+import paddle
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_multi_out_jit.py

@@ -14,14 +14,14 @@
 
 import os
 import unittest
+
 import numpy as np
+from utils import extra_cc_args, paddle_includes
 
 import paddle
-from paddle.utils.cpp_extension import load
-from paddle.utils.cpp_extension import load, get_build_directory
-from paddle.utils.cpp_extension.extension_utils import run_cmd
-from utils import paddle_includes, extra_cc_args
 from paddle.fluid.framework import _test_eager_guard
+from paddle.utils.cpp_extension import get_build_directory, load
+from paddle.utils.cpp_extension.extension_utils import run_cmd
 
 # Because Windows don't use docker, the shared lib already exists in the
 # cache dir, it will not be compiled again unless the shared lib is removed.

python/paddle/fluid/tests/custom_op/test_sysconfig.py

@@ -14,6 +14,7 @@
 
 import os
 import unittest
+
 import paddle
 
 

python/paddle/fluid/tests/custom_op/utils.py

@@ -15,6 +15,7 @@
 import os
 import sys
 from distutils.sysconfig import get_python_lib
+
 from paddle.utils.cpp_extension.extension_utils import IS_WINDOWS
 
 IS_MAC = sys.platform.startswith('darwin')

python/paddle/fluid/tests/custom_runtime/process_group_xccl.py

@@ -12,15 +12,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import random
+import unittest
+
 import numpy as np
 
 import paddle
-from paddle.fluid import core
 import paddle.fluid.core as core
-from paddle.fluid.framework import _test_eager_guard
+from paddle.fluid import core
 from paddle.fluid.dygraph.parallel import ParallelEnv
+from paddle.fluid.framework import _test_eager_guard
 
 
 def init_process_group(strategy=None):

python/paddle/fluid/tests/custom_runtime/test_collective_process_group_xccl.py

@@ -12,12 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import os
 import copy
+import os
 import subprocess
-import time
 import tempfile
+import time
+import unittest
 
 
 def start_local_trainers(
@@ -29,10 +29,10 @@ def start_local_trainers(
     log_dir=None,
 ):
     from paddle.distributed.utils.launch_utils import (  # noqa: F401
+        TrainerProc,
         find_free_ports,
-        watch_local_trainers,
         get_cluster,
-        TrainerProc,
+        watch_local_trainers,
     )
 
     current_env = copy.copy(os.environ.copy())
@@ -89,10 +89,10 @@ def start_local_trainers(
 
 def get_cluster_from_args(selected_gpus):
     from paddle.distributed.utils.launch_utils import (  # noqa: F401
+        TrainerProc,
         find_free_ports,
-        watch_local_trainers,
         get_cluster,
-        TrainerProc,
+        watch_local_trainers,
     )
 
     cluster_node_ips = '127.0.0.1'
@@ -117,10 +117,10 @@ def get_cluster_from_args(selected_gpus):
 class TestMultipleCustomCPU(unittest.TestCase):
     def run_mnist_2custom_cpu(self, target_file_name, eager_mode=True):
         from paddle.distributed.utils.launch_utils import (  # noqa: F401
+            TrainerProc,
             find_free_ports,
-            watch_local_trainers,
             get_cluster,
-            TrainerProc,
+            watch_local_trainers,
         )
 
         selected_devices = [0, 1]
@@ -179,10 +179,10 @@ class TestProcessGroup(TestMultipleCustomCPU):
 
     def test_process_group_xccl(self):
         from paddle.distributed.utils.launch_utils import (  # noqa: F401
+            TrainerProc,
             find_free_ports,
-            watch_local_trainers,
             get_cluster,
-            TrainerProc,
+            watch_local_trainers,
         )
 
         self.run_mnist_2custom_cpu('process_group_xccl.py')

python/paddle/fluid/tests/custom_runtime/test_custom_cpu_plugin.py

@@ -14,9 +14,10 @@
 
 import os
 import sys
+import tempfile
 import unittest
+
 import numpy as np
-import tempfile
 
 
 class TestCustomCPUPlugin(unittest.TestCase):

python/paddle/fluid/tests/custom_runtime/test_custom_cpu_profiler_plugin.py

@@ -14,8 +14,8 @@
 
 import os
 import sys
-import unittest
 import tempfile
+import unittest
 
 
 class TestCustomCPUProfilerPlugin(unittest.TestCase):

python/paddle/fluid/tests/custom_runtime/test_custom_cpu_to_static.py

@@ -14,10 +14,11 @@
 
 import os
 import sys
+import tempfile
 import time
 import unittest
+
 import numpy as np
-import tempfile
 
 EPOCH_NUM = 1
 BATCH_SIZE = 1024

python/paddle/fluid/tests/test_data_feeder.py

@@ -12,9 +12,10 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.fluid as fluid
 import unittest
+
 import paddle
+import paddle.fluid as fluid
 
 paddle.enable_static()
 

python/paddle/fluid/tests/test_detection.py

@@ -12,17 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.fluid as fluid
-import paddle.fluid.layers as layers
-from paddle.fluid.layers import detection
-from paddle.fluid.framework import Program, program_guard
-import unittest
 import contextlib
+import unittest
+
 import numpy as np
 from unittests.test_imperative_base import new_program_scope
-from paddle.fluid.dygraph import base
-from paddle.fluid import core
+
 import paddle
+import paddle.fluid as fluid
+import paddle.fluid.layers as layers
+from paddle.fluid import core
+from paddle.fluid.dygraph import base
+from paddle.fluid.framework import Program, program_guard
+from paddle.fluid.layers import detection
 
 paddle.enable_static()
 

python/paddle/fluid/tests/test_if_else_op.py

@@ -12,19 +12,22 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import unittest
+
+import numpy as np
+
 import paddle
+import paddle.fluid as fluid
+import paddle.fluid.core as core
 import paddle.fluid.layers as layers
-from paddle.fluid.framework import Program, program_guard
 from paddle.fluid.executor import Executor
+from paddle.fluid.framework import Program, program_guard
+from paddle.fluid.layers.control_flow import (
+    ConditionalBlock,
+    merge_lod_tensor,
+    split_lod_tensor,
+)
 from paddle.fluid.optimizer import MomentumOptimizer
-import paddle.fluid.core as core
-import paddle.fluid as fluid
-from paddle.fluid.layers.control_flow import split_lod_tensor
-from paddle.fluid.layers.control_flow import merge_lod_tensor
-from paddle.fluid.layers.control_flow import ConditionalBlock
-
-import unittest
-import numpy as np
 
 paddle.enable_static()
 

python/paddle/fluid/tests/test_lod_tensor.py

@@ -12,14 +12,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import unittest
+
+import numpy as np
+
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.lod_tensor import (
     create_lod_tensor,
     create_random_int_lodtensor,
 )
-import numpy as np
-import unittest
 
 
 class TestLoDTensor(unittest.TestCase):

python/paddle/fluid/tests/test_python_operator_overriding.py

@@ -16,10 +16,10 @@ import unittest
 
 import numpy as np
 
-import paddle.fluid.layers as layers
-import paddle.fluid.framework as framework
-import paddle.fluid as fluid
 import paddle
+import paddle.fluid as fluid
+import paddle.fluid.framework as framework
+import paddle.fluid.layers as layers
 
 paddle.enable_static()
 

python/paddle/fluid/tests/test_sequential.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 
 

python/paddle/fluid/tests/unittests/ascend_group.py

@@ -13,14 +13,15 @@
 # limitations under the License.
 
 import os
+from collections import namedtuple
+
+import paddle
 import paddle.fluid as fluid
-from paddle.fluid import unique_name
 import paddle.fluid.core as core
-import paddle
-from paddle.fluid.layer_helper import LayerHelper
 from paddle.distributed import fleet
 from paddle.distributed.fleet.meta_optimizers.ascend import ascend_optimizer
-from collections import namedtuple
+from paddle.fluid import unique_name
+from paddle.fluid.layer_helper import LayerHelper
 
 Block = namedtuple('Block', ['program'])
 Loss = namedtuple('Loss', ['block'])

python/paddle/fluid/tests/unittests/asp/asp_pruning_base.py

@@ -14,11 +14,13 @@
 # limitations under the License.
 
 import unittest
+
+import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/asp/test_asp_customized_pruning.py

@@ -14,7 +14,9 @@
 # limitations under the License.
 
 import unittest
+
 import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core

python/paddle/fluid/tests/unittests/asp/test_asp_optimize_dynamic.py

@@ -14,10 +14,12 @@
 # limitations under the License.
 
 import unittest
+
+import numpy as np
+
 import paddle
 import paddle.fluid.core as core
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
 
 
 class MyLayer(paddle.nn.Layer):

python/paddle/fluid/tests/unittests/asp/test_asp_optimize_static.py

@@ -14,11 +14,13 @@
 # limitations under the License.
 
 import unittest
+
+import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/asp/test_asp_pruning_dynamic.py

@@ -14,6 +14,7 @@
 # limitations under the License.
 
 import unittest
+
 import numpy as np
 
 import paddle

python/paddle/fluid/tests/unittests/asp/test_asp_pruning_static.py

@@ -14,11 +14,13 @@
 # limitations under the License.
 
 import unittest
+
+import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/asp/test_asp_save_load.py

@@ -14,11 +14,13 @@
 # limitations under the License.
 
 import unittest
+
+import numpy as np
+
 import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
 
 
 class MyLayer(paddle.nn.Layer):

python/paddle/fluid/tests/unittests/asp/test_asp_utils.py

@@ -13,11 +13,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import threading
+import time
 import unittest
-import threading, time
-import paddle
+
 import numpy as np
 
+import paddle
+
 
 class TestASPUtils(unittest.TestCase):
     def test_get_check_method(self):

python/paddle/fluid/tests/unittests/asp/test_fleet_with_asp_dynamic.py

@@ -13,13 +13,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.distributed.fleet as fleet
+import os
 import unittest
+
+import numpy as np
+
 import paddle
+import paddle.distributed.fleet as fleet
 import paddle.fluid.core as core
-import os
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
 
 cuda_visible_devices = os.getenv('CUDA_VISIBLE_DEVICES')
 if cuda_visible_devices is None or cuda_visible_devices == "":

python/paddle/fluid/tests/unittests/asp/test_fleet_with_asp_sharding.py

@@ -13,14 +13,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.distributed.fleet as fleet
+import os
 import unittest
+
+import numpy as np
+
 import paddle
+import paddle.distributed.fleet as fleet
 import paddle.fluid as fluid
-import os
-from paddle.static import sparsity
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
+from paddle.static import sparsity
 
 cuda_visible_devices = os.getenv('CUDA_VISIBLE_DEVICES')
 if cuda_visible_devices is None or cuda_visible_devices == "":

python/paddle/fluid/tests/unittests/asp/test_fleet_with_asp_static.py

@@ -13,14 +13,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.distributed.fleet as fleet
+import os
 import unittest
+
+import numpy as np
+
 import paddle
+import paddle.distributed.fleet as fleet
 import paddle.fluid as fluid
-import os
-from paddle.static import sparsity
 from paddle.fluid.contrib.sparsity.asp import ASPHelper
-import numpy as np
+from paddle.static import sparsity
 
 cuda_visible_devices = os.getenv('CUDA_VISIBLE_DEVICES')
 if cuda_visible_devices is None or cuda_visible_devices == "":

python/paddle/fluid/tests/unittests/auto_checkpoint_utils.py

@@ -12,15 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import paddle.fluid as fluid
 import os
+import unittest
+
+import numpy as np
 
+import paddle.fluid as fluid
 import paddle.fluid.incubate.checkpoint.auto_checkpoint as acp
-from paddle.fluid.framework import program_guard
 from paddle.fluid import unique_name
-
-import numpy as np
+from paddle.fluid.framework import program_guard
 
 BATCH_NUM = 4
 BATCH_SIZE = 1

python/paddle/fluid/tests/unittests/auto_parallel/amp_pass_unittest.py

@@ -12,14 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import random
+import unittest
+
 import numpy as np
-import paddle
+from get_gpt_model import FakeDataset, generate_model
 
+import paddle
 from paddle.distributed.fleet import auto
 from paddle.fluid.dygraph.parallel import ParallelEnv
-from get_gpt_model import FakeDataset, generate_model
 
 
 def apply_pass(use_amp=False, level=None):

python/paddle/fluid/tests/unittests/auto_parallel/auto_parallel_relaunch_model.py

@@ -13,10 +13,11 @@
 # limitations under the License.
 
 import numpy as np
+
 import paddle
 import paddle.nn as nn
-import paddle.static as static
 import paddle.nn.functional as F
+import paddle.static as static
 import paddle.utils as utils
 from paddle.distributed import fleet
 from paddle.distributed.fleet import auto

python/paddle/fluid/tests/unittests/auto_parallel/auto_parallel_relaunch_with_gpt_planner.py

@@ -12,18 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle
-import paddle.static as static
-from paddle.distributed import fleet
 import sys
 
 import numpy as np
 
+import paddle
+import paddle.static as static
+from paddle.distributed import fleet
+
 sys.path.append("..")
 import auto_parallel_gpt_model as modeling
 from auto_parallel_gpt_model import (
-    GPTModel,
     GPTForPretraining,
+    GPTModel,
     GPTPretrainingCriterion,
 )
 

python/paddle/fluid/tests/unittests/auto_parallel/auto_parallel_relaunch_with_planner.py

@@ -15,16 +15,18 @@
 import paddle
 import paddle.static as static
 from paddle.distributed import fleet
-from paddle.distributed.auto_parallel.cost import CostEstimator
 from paddle.distributed.auto_parallel.cluster import Cluster
+from paddle.distributed.auto_parallel.cost import CostEstimator
 from paddle.distributed.auto_parallel.dist_context import (
     get_default_distributed_context,
 )
 
 
 def train():
-    from auto_parallel_relaunch_model import mlp_pretrain_forward
-    from auto_parallel_relaunch_model import batch_generator_creator
+    from auto_parallel_relaunch_model import (
+        batch_generator_creator,
+        mlp_pretrain_forward,
+    )
 
     dist_strategy = fleet.DistributedStrategy()
     # init parallel optimizer

python/paddle/fluid/tests/unittests/auto_parallel/clip_grad_by_global_norm.py

@@ -12,14 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import random
+import unittest
+
 import numpy as np
-import paddle
+from get_gpt_model import FakeDataset, generate_model
 
+import paddle
 from paddle.distributed.fleet import auto
 from paddle.fluid.dygraph.parallel import ParallelEnv
-from get_gpt_model import FakeDataset, generate_model
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/engine_api.py

@@ -12,17 +12,18 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
 import os
+import tempfile
+
 import numpy as np
+
 import paddle
-import paddle.static as static
-import paddle.utils as utils
 import paddle.nn as nn
 import paddle.nn.functional as F
-from paddle.io import Dataset
-
+import paddle.static as static
+import paddle.utils as utils
 from paddle.distributed.fleet import auto
+from paddle.io import Dataset
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/engine_api_dp.py

@@ -12,15 +12,16 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
 import os
+import tempfile
+
 import numpy as np
+
 import paddle
 import paddle.nn as nn
 import paddle.nn.functional as F
-from paddle.io import Dataset
-
 from paddle.distributed.fleet import auto
+from paddle.io import Dataset
 
 paddle.enable_static()
 batch_size = 2

python/paddle/fluid/tests/unittests/auto_parallel/get_gpt_model.py

@@ -12,9 +12,10 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import random
 import sys
+
 import numpy as np
-import random
 
 import paddle
 from paddle.distributed.fleet import auto
@@ -22,8 +23,8 @@ from paddle.distributed.fleet import auto
 sys.path.append("..")
 import auto_parallel_gpt_model as modeling
 from auto_parallel_gpt_model import (
-    GPTModel,
     GPTForPretraining,
+    GPTModel,
     GPTPretrainingCriterion,
 )
 

python/paddle/fluid/tests/unittests/auto_parallel/gradient_merge_pass_unittest.py

@@ -12,14 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import random
+import unittest
+
 import numpy as np
-import paddle
+from get_gpt_model import FakeDataset, generate_model
 
+import paddle
 from paddle.distributed.fleet import auto
 from paddle.fluid.dygraph.parallel import ParallelEnv
-from get_gpt_model import FakeDataset, generate_model
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/high_order_grad.py

@@ -12,8 +12,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle
 import numpy as np
+
+import paddle
 from paddle.distributed.fleet import auto
 from paddle.incubate.autograd import Hessian
 

python/paddle/fluid/tests/unittests/auto_parallel/iterable_dataset.py

@@ -12,13 +12,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
 import os
+import tempfile
+
 import numpy as np
+
 import paddle
 import paddle.nn as nn
 import paddle.nn.functional as F
-
 from paddle.distributed.fleet import auto
 
 paddle.enable_static()

python/paddle/fluid/tests/unittests/auto_parallel/launch.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import os
+
 from paddle.distributed.fleet import launch
 from paddle.distributed.fleet.launch_utils import run_with_coverage
 

python/paddle/fluid/tests/unittests/auto_parallel/optimization_tuner_api.py

@@ -12,12 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from engine_api_dp import MyDataset
+
 import paddle
 import paddle.nn as nn
 import paddle.nn.functional as F
-
 from paddle.distributed.fleet import auto
-from engine_api_dp import MyDataset
 
 paddle.enable_static()
 batch_size = 16

python/paddle/fluid/tests/unittests/auto_parallel/recompute_pass_unittest.py

@@ -12,14 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import random
+import unittest
+
 import numpy as np
-import paddle
+from get_gpt_model import FakeDataset, generate_model
 
+import paddle
 from paddle.distributed.fleet import auto
 from paddle.fluid.dygraph.parallel import ParallelEnv
-from get_gpt_model import FakeDataset, generate_model
 
 
 def apply_pass(use_recompute=False, no_recompute_segments=[]):

python/paddle/fluid/tests/unittests/auto_parallel/sharding_pass_unittest.py

@@ -12,14 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import random
+import unittest
+
 import numpy as np
-import paddle
+from get_gpt_model import FakeDataset, generate_model
 
+import paddle
 from paddle.distributed.fleet import auto
 from paddle.fluid.dygraph.parallel import ParallelEnv
-from get_gpt_model import FakeDataset, generate_model
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_auto_parallel_relaunch.py

@@ -12,12 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
-import unittest
-import os
-import sys
 import json
+import os
 import subprocess
+import sys
+import tempfile
+import unittest
 
 cluster_json = """
 {

python/paddle/fluid/tests/unittests/auto_parallel/test_base_cost.py

@@ -12,38 +12,35 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import os
 import json
+import os
 import tempfile
+import unittest
+
+from test_cluster import cluster_json
 
 import paddle
 import paddle.nn as nn
-import paddle.static as static
 import paddle.nn.functional as F
+import paddle.static as static
 import paddle.utils as utils
-from paddle.distributed.fleet import auto
-from paddle.distributed.auto_parallel.completion import Completer
-from paddle.distributed.auto_parallel.dist_context import DistributedContext
 from paddle.distributed import fleet
-from paddle.distributed.auto_parallel.parallelizer import AutoParallelizer
 from paddle.distributed.auto_parallel.cluster import Cluster
-from paddle.distributed.auto_parallel.cost.base_cost import (
-    build_comp_desc_from_dist_op,
-)
-from paddle.distributed.auto_parallel.cost.base_cost import (
-    build_comm_desc_from_dist_op,
+from paddle.distributed.auto_parallel.completion import Completer
+from paddle.distributed.auto_parallel.cost import (
+    AllreduceSumOpCost,
+    _g_op_cost_factory,
 )
 from paddle.distributed.auto_parallel.cost.base_cost import (
     build_comm_costs_from_descs,
-)
-from paddle.distributed.auto_parallel.cost.base_cost import (
+    build_comm_desc_from_dist_op,
     build_comp_costs_from_descs,
+    build_comp_desc_from_dist_op,
+    build_dp_costs,
 )
-from paddle.distributed.auto_parallel.cost.base_cost import build_dp_costs
-from paddle.distributed.auto_parallel.cost import AllreduceSumOpCost
-from paddle.distributed.auto_parallel.cost import _g_op_cost_factory
-from test_cluster import cluster_json
+from paddle.distributed.auto_parallel.dist_context import DistributedContext
+from paddle.distributed.auto_parallel.parallelizer import AutoParallelizer
+from paddle.distributed.fleet import auto
 
 paddle.enable_static()
 _global_parallel_strategy = "dp_mp_pp"

python/paddle/fluid/tests/unittests/auto_parallel/test_cluster.py

@@ -12,13 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import json
+import os
 import tempfile
 import unittest
-import os
-import json
 
-from paddle.distributed.auto_parallel.cluster import Cluster
-from paddle.distributed.auto_parallel.cluster import get_default_cluster
+from paddle.distributed.auto_parallel.cluster import (
+    Cluster,
+    get_default_cluster,
+)
 
 cluster_json = """
 {

python/paddle/fluid/tests/unittests/auto_parallel/test_cluster_v2.py

@@ -13,9 +13,8 @@
 # limitations under the License
 
 import unittest
-from paddle.distributed.auto_parallel.cluster_v2 import Device
-from paddle.distributed.auto_parallel.cluster_v2 import Link
-from paddle.distributed.auto_parallel.cluster_v2 import DeviceMesh
+
+from paddle.distributed.auto_parallel.cluster_v2 import Device, DeviceMesh, Link
 
 
 class TestDeviceMesh(unittest.TestCase):

python/paddle/fluid/tests/unittests/auto_parallel/test_comm_cost.py

@@ -12,23 +12,25 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import os
 import json
+import os
 import tempfile
+import unittest
+
+from test_cluster import cluster_json, multi_cluster_json
 
 import paddle
 from paddle.distributed.auto_parallel.cluster import Cluster
-from paddle.distributed.auto_parallel.cost import CommContext
-from paddle.distributed.auto_parallel.cost import build_comm_desc
-from paddle.distributed.auto_parallel.cost import AllreduceSumOpCost
-from paddle.distributed.auto_parallel.cost import AllgatherOpCost
-from paddle.distributed.auto_parallel.cost import BroadcastOpCost
-from paddle.distributed.auto_parallel.cost import SendOpCost
-from paddle.distributed.auto_parallel.cost import RecvOpCost
-from paddle.distributed.auto_parallel.cost import IdentityOpCost
-
-from test_cluster import cluster_json, multi_cluster_json
+from paddle.distributed.auto_parallel.cost import (
+    AllgatherOpCost,
+    AllreduceSumOpCost,
+    BroadcastOpCost,
+    CommContext,
+    IdentityOpCost,
+    RecvOpCost,
+    SendOpCost,
+    build_comm_desc,
+)
 
 
 class TestCommOpCost(unittest.TestCase):

python/paddle/fluid/tests/unittests/auto_parallel/test_comp_cost.py

@@ -12,131 +12,81 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import os
 import json
+import os
+import unittest
+
+from test_cluster import cluster_json
 
 from paddle.distributed.auto_parallel.cluster import Cluster
-from paddle.distributed.auto_parallel.cost.comp_op_cost import AssignOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import AssignValueOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import BeamSearchOpCost
 from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    AssignOpCost,
+    AssignValueOpCost,
     BeamSearchDecodeOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import CastOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import ConcatOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    ElementwiseAddOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    BeamSearchOpCost,
+    CastOpCost,
+    ConcatOpCost,
+    DropoutGradOpCost,
     ElementwiseAddGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    ElementwiseDivOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    ElementwiseAddOpCost,
     ElementwiseDivGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    ElementwiseMulOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    ElementwiseDivOpCost,
     ElementwiseMulGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    ElementwiseMulOpCost,
     ElementwiseSubOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import EmbeddingOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
     EmbeddingGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    FillConstantOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    EmbeddingOpCost,
     FillConstantBatchSizeLikeOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import GatherOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import GeluOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import GeluGradOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    FillConstantOpCost,
+    FusedSoftmaxMaskUpperTriangleGradOpCost,
+    FusedSoftmaxMaskUpperTriangleOpCost,
+    GatherOpCost,
+    GeluGradOpCost,
+    GeluOpCost,
     GreaterEqualOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import IncrementOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import IsEmptyOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import LayerNormOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    IncrementOpCost,
+    IsEmptyOpCost,
     LayerNormGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import LessThanOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import LogicalNotOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import LogicalAndOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import LodResetOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import LogOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    LookupTableV2OpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    LayerNormOpCost,
+    LessThanOpCost,
+    LodResetOpCost,
+    LogicalAndOpCost,
+    LogicalNotOpCost,
+    LogOpCost,
     LookupTableV2GradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import MatmulOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import MatmulV2OpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    LookupTableV2OpCost,
+    MatmulOpCost,
     MatmulV2GradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import MemcpyOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import MulOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import MulGradOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import OneHotOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    MatmulV2OpCost,
+    MemcpyOpCost,
+    MulGradOpCost,
+    MulOpCost,
+    OneHotOpCost,
     ReadFromArrayOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import ReduceSumOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    ReduceMeanGradOpCost,
+    ReduceMeanOpCost,
     ReduceSumGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import Reshape2OpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    ReduceSumOpCost,
     Reshape2GradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import ReduceMeanOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    ReduceMeanGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SamplingIdOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import ScaleOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SliceOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SoftmaxOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SoftmaxGradOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    SoftmaxWithCrossEntropyOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    Reshape2OpCost,
+    SamplingIdOpCost,
+    ScaleOpCost,
+    SliceOpCost,
+    SoftmaxGradOpCost,
+    SoftmaxOpCost,
     SoftmaxWithCrossEntropyGradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SplitOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import Squeeze2OpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SquareOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SquareGradOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import SumOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import TopKOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import Transpose2OpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    SoftmaxWithCrossEntropyOpCost,
+    SplitOpCost,
+    SquareGradOpCost,
+    SquareOpCost,
+    Squeeze2OpCost,
+    SumOpCost,
+    TopKOpCost,
     Transpose2GradOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import Unsqueeze2OpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
+    Transpose2OpCost,
+    Unsqueeze2OpCost,
     WriteToArrayOpCost,
 )
-from paddle.distributed.auto_parallel.cost.comp_op_cost import DropoutGradOpCost
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    FusedSoftmaxMaskUpperTriangleOpCost,
-)
-from paddle.distributed.auto_parallel.cost.comp_op_cost import (
-    FusedSoftmaxMaskUpperTriangleGradOpCost,
-)
-
-from test_cluster import cluster_json
 
 
 class TestCompOpCost(unittest.TestCase):

python/paddle/fluid/tests/unittests/auto_parallel/test_conditional_block_reshard.py

@@ -17,9 +17,8 @@ import unittest
 import paddle
 import paddle.nn as nn
 import paddle.nn.functional as F
-
-from paddle.static import InputSpec
 from paddle.distributed.fleet import auto
+from paddle.static import InputSpec
 
 
 class MLPLayer(nn.Layer):

python/paddle/fluid/tests/unittests/auto_parallel/test_converter.py

@@ -12,11 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
-import unittest
 import os
-import sys
 import subprocess
+import sys
+import tempfile
+import unittest
+
 from paddle.distributed.auto_parallel.converter import Converter
 
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_assign.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
 
@@ -38,8 +39,8 @@ def make_program():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_attr_v2.py

@@ -12,31 +12,29 @@
 # See the License for the specific language governing permissions and
 # limitations under the License
 
-import unittest
 import copy
+import unittest
 
-import paddle
 import numpy as np
+
+import paddle
 import paddle.nn as nn
-import paddle.static as static
 import paddle.nn.functional as F
-
+import paddle.static as static
 from paddle.distributed import fleet
-from paddle.distributed.fleet import auto
 from paddle.distributed.auto_parallel.dist_context import (
     DistributedContext,
     set_default_distributed_context,
 )
+from paddle.distributed.auto_parallel.process_mesh_v2 import ProcessMesh
 from paddle.distributed.auto_parallel.utils import (
-    _copy_dist_attr_to_cpp,
     _copy_dist_attr_from_cpp,
-    _copy_dist_attr_to_cpp_for_graph,
     _copy_dist_attr_from_cpp_for_graph,
+    _copy_dist_attr_to_cpp,
+    _copy_dist_attr_to_cpp_for_graph,
 )
-
-from paddle.fluid.core import TensorDistAttr
-from paddle.fluid.core import OperatorDistAttr
-from paddle.distributed.auto_parallel.process_mesh_v2 import ProcessMesh
+from paddle.distributed.fleet import auto
+from paddle.fluid.core import OperatorDistAttr, TensorDistAttr
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_context.py

@@ -12,18 +12,18 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import copy
+import unittest
 
-import paddle
 import numpy as np
+
+import paddle
 import paddle.nn as nn
-import paddle.static as static
 import paddle.nn.functional as F
-
+import paddle.static as static
 from paddle.distributed import fleet
-from paddle.distributed.fleet import auto
 from paddle.distributed.auto_parallel.dist_context import DistributedContext
+from paddle.distributed.fleet import auto
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_embedding.py

@@ -13,11 +13,12 @@
 # limitations under the License.
 
 import unittest
-import paddle
-from paddle.distributed.fleet import auto
 
 from test_dist_pnorm import parallelizer
 
+import paddle
+from paddle.distributed.fleet import auto
+
 paddle.enable_static()
 
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_matmul.py

@@ -13,9 +13,9 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
-
 from paddle.fluid import program_guard
 from paddle.fluid.backward import append_backward
 
@@ -106,8 +106,8 @@ def matmulv2_dp2mp2(init_x, init_y, trans_x, trans_y):
 
 def parallelizer(program_func, *args, **kwargs):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program, loss = program_func(*args, **kwargs)
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_op_cost.py

@@ -12,21 +12,18 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import copy
+import unittest
 
 import paddle
-
-from paddle.distributed.fleet import auto
 from paddle.distributed.auto_parallel.cluster import Cluster
 from paddle.distributed.auto_parallel.operators.common import (
     get_distributed_operator_impl_container,
     is_elementwise_op,
 )
-
+from paddle.distributed.fleet import auto
 from paddle.fluid import program_guard
 from paddle.fluid.backward import append_backward
-from paddle.fluid.backward import append_backward
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_pnorm.py

@@ -13,9 +13,9 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
-
 from paddle.fluid import program_guard
 from paddle.fluid.backward import append_backward
 
@@ -50,8 +50,8 @@ def make_program_serial():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program, loss = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_reshape.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
 
@@ -37,8 +38,8 @@ def make_program_dp2():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_scale.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
 
@@ -34,8 +35,8 @@ def make_program():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_shape.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
 
@@ -34,8 +35,8 @@ def make_program():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_slice.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
 
@@ -56,8 +57,8 @@ def make_program_serial():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_dist_split.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import unittest
+
 import paddle
 from paddle.distributed.fleet import auto
 
@@ -34,8 +35,8 @@ def make_program_dp2():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_engine_api.py

@@ -12,11 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
-import unittest
 import os
-import sys
 import subprocess
+import sys
+import tempfile
+import unittest
 
 
 class TestEngineAPI(unittest.TestCase):

python/paddle/fluid/tests/unittests/auto_parallel/test_engine_api_dp.py

@@ -12,11 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
-import unittest
 import os
-import sys
 import subprocess
+import sys
+import tempfile
+import unittest
 
 
 class TestEngineAPI(unittest.TestCase):

python/paddle/fluid/tests/unittests/auto_parallel/test_engine_api_error.py

@@ -13,14 +13,15 @@
 # limitations under the License.
 
 import unittest
+
 import numpy as np
+
 import paddle
-import paddle.static as static
 import paddle.nn as nn
 import paddle.nn.functional as F
-from paddle.io import Dataset
-
+import paddle.static as static
 from paddle.distributed.fleet import auto
+from paddle.io import Dataset
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_engine_callbacks.py

@@ -12,20 +12,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-import tempfile
+import random
 import shutil
+import tempfile
 import time
-import random
+import unittest
 
 import paddle
 import paddle.vision.transforms as T
-
-from paddle.static import InputSpec
-from paddle.distributed.fleet import auto
 from paddle.distributed.auto_parallel.callbacks import config_callbacks
-from paddle.vision.models import LeNet
+from paddle.distributed.fleet import auto
+from paddle.static import InputSpec
 from paddle.vision.datasets import MNIST
+from paddle.vision.models import LeNet
 
 paddle.enable_static()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_fp16_assign.py

@@ -12,8 +12,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
 import copy
+import unittest
 
 import paddle
 from paddle.distributed.fleet import auto
@@ -65,8 +65,8 @@ def make_program():
 
 def parallelizer(program_func, rank):
     from paddle.distributed.auto_parallel.completion import Completer
-    from paddle.distributed.auto_parallel.partitioner import Partitioner
     from paddle.distributed.auto_parallel.dist_context import DistributedContext
+    from paddle.distributed.auto_parallel.partitioner import Partitioner
 
     main_program, start_program = program_func()
 

python/paddle/fluid/tests/unittests/auto_parallel/test_group_operators.py

@@ -14,6 +14,7 @@
 
 import sys
 import unittest
+
 import numpy as np
 
 import paddle
@@ -22,8 +23,8 @@ import paddle.static as static
 sys.path.append("..")
 import auto_parallel_gpt_model as modeling
 from auto_parallel_gpt_model import (
-    GPTModel,
     GPTForPretraining,
+    GPTModel,
     GPTPretrainingCriterion,
 )
 
@@ -111,12 +112,12 @@ class TestGroupOperators(unittest.TestCase):
             sequence_len,
             vocab_size,
         )
-        from paddle.distributed.auto_parallel.tuner.rule_based_tuner import (
-            RuleBasedTuner,
-        )
         from paddle.distributed.auto_parallel.dist_context import (
             DistributedContext,
         )
+        from paddle.distributed.auto_parallel.tuner.rule_based_tuner import (
+            RuleBasedTuner,
+        )
 
         dist_context = DistributedContext()
         tuner = RuleBasedTuner(dist_context)

python/paddle/fluid/tests/unittests/auto_parallel/test_high_order_grad.py

@@ -12,11 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import tempfile
-import unittest
 import os
-import sys
 import subprocess
+import sys
+import tempfile
+import unittest
 
 
 class TestHighOrderGrad(unittest.TestCase):