diff --git a/paddle/infrt/tests/models/efficientnet-b4/net/utils.py b/paddle/infrt/tests/models/efficientnet-b4/net/utils.py
index 9a060306995bba614548d29d18e72eca6d42b834..0617870bf019097ca6e0acc68eb95e08c67e2da3 100644
--- a/paddle/infrt/tests/models/efficientnet-b4/net/utils.py
+++ b/paddle/infrt/tests/models/efficientnet-b4/net/utils.py
@@ -239,7 +239,7 @@ def efficientnet_params(model_name):
     return params_dict[model_name]
 
 
-class BlockDecoder(object):
+class BlockDecoder:
     """Block Decoder for readability, straight from the official TensorFlow repository"""
 
     @staticmethod
diff --git a/paddle/phi/api/yaml/generator/api_base.py b/paddle/phi/api/yaml/generator/api_base.py
index 53b950b63f05231b0e8d7067b79407aea6158ca1..696ad8736b90e4224c0bb770f3b7702da90d432f 100644
--- a/paddle/phi/api/yaml/generator/api_base.py
+++ b/paddle/phi/api/yaml/generator/api_base.py
@@ -19,7 +19,7 @@ PREFIX_TENSOR_NAME = 'input_'
 PREFIX_META_TENSOR_NAME = 'meta_'
 
 
-class BaseAPI(object):
+class BaseAPI:
     def __init__(self, api_item_yaml):
         self.api = self.get_api_name(api_item_yaml)
 
diff --git a/python/paddle/audio/functional/window.py b/python/paddle/audio/functional/window.py
index 472c56b87acf95db0dc0c04f487af90caa5a3bb4..4836afbb61d03f81422b8102efbd63a00b7b794e 100644
--- a/python/paddle/audio/functional/window.py
+++ b/python/paddle/audio/functional/window.py
@@ -19,7 +19,7 @@ import paddle
 from paddle import Tensor
 
 
-class WindowFunctionRegister(object):
+class WindowFunctionRegister:
     def __init__(self):
         self._functions_dict = dict()
 
diff --git a/python/paddle/autograd/py_layer.py b/python/paddle/autograd/py_layer.py
index 2efb1ece83e4366a9b65663f51e1ba73bfbe64dc..252cfd5d91decc73ee6c9b40f4fa542d70e47e67 100644
--- a/python/paddle/autograd/py_layer.py
+++ b/python/paddle/autograd/py_layer.py
@@ -21,7 +21,7 @@ from paddle.fluid import core
 __all__ = []
 
 
-class LegacyPyLayerContext(object):
+class LegacyPyLayerContext:
     """
     The object of this class is a context that is used in PyLayer to enhance the function.
 
@@ -131,7 +131,7 @@ def with_mateclass(meta, *bases):
     return type.__new__(impl, "impl", (), {})
 
 
-class CPyLayer(object):
+class CPyLayer:
     @classmethod
     @dygraph_only
     def apply(cls, *args, **kwargs):
@@ -336,7 +336,7 @@ class LegacyPyLayer(with_mateclass(LayerMeta, CPyLayer)):
         )
 
 
-class EagerPyLayerContext(object):
+class EagerPyLayerContext:
     def save_for_backward(self, *tensors):
         """
         Saves given tensors that backward need. Use ``saved_tensor`` in the `backward` to get the saved tensors.
diff --git a/python/paddle/dataset/imikolov.py b/python/paddle/dataset/imikolov.py
index 4630d88e21a7cb8873df699c982d36e24c3a72bf..f618bbc2c4867c2e99b3813df987d4a96739abce 100644
--- a/python/paddle/dataset/imikolov.py
+++ b/python/paddle/dataset/imikolov.py
@@ -31,7 +31,7 @@ URL = 'https://dataset.bj.bcebos.com/imikolov%2Fsimple-examples.tgz'
 MD5 = '30177ea32e27c525793142b6bf2c8e2d'
 
 
-class DataType(object):
+class DataType:
     NGRAM = 1
     SEQ = 2
 
diff --git a/python/paddle/dataset/movielens.py b/python/paddle/dataset/movielens.py
index fd57ad8edf758ffab672296507177d337815fc75..7a47293c3c7e0421e03e66c387d2c85c5a4da7c3 100644
--- a/python/paddle/dataset/movielens.py
+++ b/python/paddle/dataset/movielens.py
@@ -38,7 +38,7 @@ URL = 'https://dataset.bj.bcebos.com/movielens%2Fml-1m.zip'
 MD5 = 'c4d9eecfca2ab87c1945afe126590906'
 
 
-class MovieInfo(object):
+class MovieInfo:
     """
     Movie id, title and categories information are stored in MovieInfo.
     """
@@ -69,7 +69,7 @@ class MovieInfo(object):
         return self.__str__()
 
 
-class UserInfo(object):
+class UserInfo:
     """
     User id, gender, age, and job information are stored in UserInfo.
     """
diff --git a/python/paddle/distributed/auto_parallel/cluster_v2.py b/python/paddle/distributed/auto_parallel/cluster_v2.py
index debcb078f6eb444809c4c621ca8dd413902ab216..1ec2332ad4003bd2d1dad8c19203f8fbac80741f 100644
--- a/python/paddle/distributed/auto_parallel/cluster_v2.py
+++ b/python/paddle/distributed/auto_parallel/cluster_v2.py
@@ -116,7 +116,7 @@ class DeviceMesh(core.DeviceMesh):
         return self._mesh
 
 
-# class Cluster(object):
+# class Cluster:
 #     """
 #     The cluster represents the hardware resource.
 #     """
diff --git a/python/paddle/distributed/auto_parallel/converter.py b/python/paddle/distributed/auto_parallel/converter.py
index d0fae414b1981d291c8b31f99dfb232e289b31e0..cc0966be4aba4c06b0288b409ad35c649e0bad6e 100644
--- a/python/paddle/distributed/auto_parallel/converter.py
+++ b/python/paddle/distributed/auto_parallel/converter.py
@@ -19,7 +19,7 @@ import numpy as np
 from ..utils.log_utils import get_logger
 
 
-class Converter(object):
+class Converter:
     """
     Converter is a class object for auto parallel to convert tensors from
     one parallel strategy to another one. Tensors will merge and slice value
diff --git a/python/paddle/distributed/auto_parallel/cost_model.py b/python/paddle/distributed/auto_parallel/cost_model.py
index f335244e36073ecbd945a104b28fa2ce1ca85d3e..73e899614d433201be7b1d741658fdd0a2b5f8f2 100644
--- a/python/paddle/distributed/auto_parallel/cost_model.py
+++ b/python/paddle/distributed/auto_parallel/cost_model.py
@@ -35,7 +35,7 @@ class CostNodeType(Enum):
     NOP = 5
 
 
-class Cost(object):
+class Cost:
     def __init__(self):
         self.runtime = None
         self.static_mem = None
@@ -49,7 +49,7 @@ class CostModelMode(Enum):
     MIXED = 3
 
 
-class CostNode(object):
+class CostNode:
     def __init__(self, node, node_type, id=None):
         self.id = id
         self.node = node
@@ -172,7 +172,7 @@ class CompOpCostNode(CostNode):
             self.cost = 0.0
 
 
-class PipeEvent(object):
+class PipeEvent:
     def __init__(self, stage_id, event_name, duration, start_time=-1):
         self.stage_id = stage_id
         self.name = event_name
@@ -181,7 +181,7 @@ class PipeEvent(object):
         self.e_time = -1
 
 
-class CostModel(object):
+class CostModel:
     def __init__(
         self,
         mode=CostModelMode.BENCHMARKING,
diff --git a/python/paddle/distributed/auto_parallel/dist_context.py b/python/paddle/distributed/auto_parallel/dist_context.py
index 4b4fca8730cd9a13b0f4ba2b6f90e8b287e0fcb9..f410468f45b8584c8152c317cdaf4977a93df939 100644
--- a/python/paddle/distributed/auto_parallel/dist_context.py
+++ b/python/paddle/distributed/auto_parallel/dist_context.py
@@ -1146,7 +1146,7 @@ class DistributedOperatorContext:
         return kinputs, koutputs
 
 
-class BlockState(object):
+class BlockState:
     def __init__(self):
         self.nblock = 0
         self.forward_indices = []
diff --git a/python/paddle/distributed/auto_parallel/helper.py b/python/paddle/distributed/auto_parallel/helper.py
index 3c13f3d9abc69ee7959a143109de94eb0577c34b..31deaea4275eb969a5e9bed344663cd78b054236 100644
--- a/python/paddle/distributed/auto_parallel/helper.py
+++ b/python/paddle/distributed/auto_parallel/helper.py
@@ -192,7 +192,7 @@ class BuildInfo:
         self.states = defaultdict(bool)
 
 
-class ProgramHelper(object):
+class ProgramHelper:
     """
     A Helper class for Engine to provides different Program IR according specified 'mode'.
     """
diff --git a/python/paddle/distributed/auto_parallel/interface.py b/python/paddle/distributed/auto_parallel/interface.py
index 124f622d40f049f224d3405379b76d625e664f33..d2f7e894149c79ba2a9061e021be3bacf6491dd7 100644
--- a/python/paddle/distributed/auto_parallel/interface.py
+++ b/python/paddle/distributed/auto_parallel/interface.py
@@ -220,7 +220,7 @@ def recompute(op):
 _g_collections = {}
 
 
-class CollectionNames(object):
+class CollectionNames:
     FETCHES = "fetches"
     LOGGING = "logging"
 
diff --git a/python/paddle/distributed/auto_parallel/partitioner.py b/python/paddle/distributed/auto_parallel/partitioner.py
index 2a7b7f3e67daabfd017ca7563294cccc23c0598a..6ec52ff69796fe183eccfb12e7b9040df194f16e 100644
--- a/python/paddle/distributed/auto_parallel/partitioner.py
+++ b/python/paddle/distributed/auto_parallel/partitioner.py
@@ -32,7 +32,7 @@ __not_shape_var_type__ = [
 ]
 
 
-class Partitioner(object):
+class Partitioner:
     """
     warning:: Partitioner is experimental and subject to change.
 
diff --git a/python/paddle/distributed/auto_parallel/process_mesh.py b/python/paddle/distributed/auto_parallel/process_mesh.py
index 9d8a44b92d2e2b964c0c1f76faff865b211c0785..1630289dde8aabaafa05125c2df0d39544a023fd 100644
--- a/python/paddle/distributed/auto_parallel/process_mesh.py
+++ b/python/paddle/distributed/auto_parallel/process_mesh.py
@@ -39,7 +39,7 @@ def reset_current_process_mesh():
     _g_current_process_mesh = _g_previous_process_mesh
 
 
-class ProcessMesh(object):
+class ProcessMesh:
     """
     The `Processmesh` object describes the topology of the used processes.
 
diff --git a/python/paddle/distributed/auto_parallel/strategy.py b/python/paddle/distributed/auto_parallel/strategy.py
index 12b6a9b78a3cdbba13647ccc7aa1b7e826e0df90..dcfd453f63a33a42a090e596f9283e6bfbf0538b 100644
--- a/python/paddle/distributed/auto_parallel/strategy.py
+++ b/python/paddle/distributed/auto_parallel/strategy.py
@@ -16,7 +16,7 @@ import copy
 from . import constants
 
 
-class BaseConfig(object):
+class BaseConfig:
     def __init__(self, category, config_dict=None):
         self._category = category
         self._config_dict = None
diff --git a/python/paddle/distributed/auto_parallel/tuner/config.py b/python/paddle/distributed/auto_parallel/tuner/config.py
index 4ed6340eccb6f25e6c5e6c0921e517c7adc4d336..7bb9d4f18bcef00748f591513e00415f88172cef 100644
--- a/python/paddle/distributed/auto_parallel/tuner/config.py
+++ b/python/paddle/distributed/auto_parallel/tuner/config.py
@@ -25,7 +25,7 @@ def _get_pass_config(strategy, pass_name):
     return config
 
 
-class TuningConfig(object):
+class TuningConfig:
     """
     A uniform config wrap:
     distributed strategy: the user defined configuration for optimization pass
diff --git a/python/paddle/distributed/auto_parallel/tuner/recorder.py b/python/paddle/distributed/auto_parallel/tuner/recorder.py
index 2c838cfb1492164fbb0e27f6085238f8271dfdcb..8174ba22cf78bfc24ab8fcc9987f30f041477d2e 100644
--- a/python/paddle/distributed/auto_parallel/tuner/recorder.py
+++ b/python/paddle/distributed/auto_parallel/tuner/recorder.py
@@ -18,7 +18,7 @@
 import numpy as np
 
 
-class MetricRecord(object):
+class MetricRecord:
     """
     One record for a single metric at a given execution step.
     """
@@ -62,7 +62,7 @@ class MetricRecord(object):
         return "MetricRecord(value={}, step={})".format(self.value, self.step)
 
 
-class MetricRecords(object):
+class MetricRecords:
     """
     Records of a single metric across different executions.
     """
@@ -143,7 +143,7 @@ class MetricRecords(object):
         return records
 
 
-class MetricsRecorder(object):
+class MetricsRecorder:
     """
     Record the values for all metrics.
     """
diff --git a/python/paddle/distributed/auto_parallel/tuner/storable.py b/python/paddle/distributed/auto_parallel/tuner/storable.py
index fb03070ad099a7b122f514c3c56386a35f3cbe92..01e10b4a3b4965261ace9918c350faa96525758a 100644
--- a/python/paddle/distributed/auto_parallel/tuner/storable.py
+++ b/python/paddle/distributed/auto_parallel/tuner/storable.py
@@ -18,7 +18,7 @@
 import json
 
 
-class Storable(object):
+class Storable:
     def get_state(self):
         raise NotImplementedError
 
diff --git a/python/paddle/distributed/auto_parallel/tuner/tunable_space.py b/python/paddle/distributed/auto_parallel/tuner/tunable_space.py
index e3e503401b45371d9372aec6b6ff4ee9d42940df..a4383f5385dee34918684dcfc17b5a2dc91ef90d 100644
--- a/python/paddle/distributed/auto_parallel/tuner/tunable_space.py
+++ b/python/paddle/distributed/auto_parallel/tuner/tunable_space.py
@@ -22,7 +22,7 @@ from .tunable_variable import IntRange
 from .tunable_variable import FloatRange
 
 
-class TunableSpace(object):
+class TunableSpace:
     """
     A TunableSpace is constructed by the tunable variables.
     """
diff --git a/python/paddle/distributed/auto_parallel/tuner/tunable_variable.py b/python/paddle/distributed/auto_parallel/tuner/tunable_variable.py
index 74594a8b4ad6dcbeafc0905b202b64dd2858fa83..3f45c68c1de6db8cc4d9bc7f9f7a79ce1f6e9037 100644
--- a/python/paddle/distributed/auto_parallel/tuner/tunable_variable.py
+++ b/python/paddle/distributed/auto_parallel/tuner/tunable_variable.py
@@ -18,7 +18,7 @@
 import numpy as np
 
 
-class TunableVariable(object):
+class TunableVariable:
     """
     Tunablevariable base class.
     """
diff --git a/python/paddle/distributed/communication/batch_isend_irecv.py b/python/paddle/distributed/communication/batch_isend_irecv.py
index 073ccb0b41e131f6ac74176c38eae4d9fa56614f..d3f0372b6853327d9573584b6c0047eb203db55d 100644
--- a/python/paddle/distributed/communication/batch_isend_irecv.py
+++ b/python/paddle/distributed/communication/batch_isend_irecv.py
@@ -22,7 +22,7 @@ from paddle.distributed.communication.group import (
 )
 
 
-class P2POp(object):
+class P2POp:
     """
     A class that makes point-to-point operations for "batch_isend_irecv".
 
diff --git a/python/paddle/distributed/elastic.py b/python/paddle/distributed/elastic.py
index 55b73ab315bb4f3e96045e6de6e310c3960bf931..082fdd3c07bea54ef9e9f7831856e7cf11d14d76 100644
--- a/python/paddle/distributed/elastic.py
+++ b/python/paddle/distributed/elastic.py
@@ -16,7 +16,7 @@ import argparse
 import os
 
 
-class Command(object):
+class Command:
     def __init__(self, server, name):
         import etcd3
 
diff --git a/python/paddle/distributed/entry_attr.py b/python/paddle/distributed/entry_attr.py
index be54d4ab7b17ab3bb2bf442cf9c5782a3bf8c8eb..dcd5153bb5f1ea85533595838e5ed8912f53aeaf 100644
--- a/python/paddle/distributed/entry_attr.py
+++ b/python/paddle/distributed/entry_attr.py
@@ -15,7 +15,7 @@
 __all__ = []
 
 
-class EntryAttr(object):
+class EntryAttr:
     """
     Entry Config for paddle.static.nn.sparse_embedding with Parameter Server.
 
diff --git a/python/paddle/distributed/fleet/base/distributed_strategy.py b/python/paddle/distributed/fleet/base/distributed_strategy.py
index c32b1f2d68c59383db82bd908ca9ef507b0664fc..5b3b599513d22d191c75d76cf5f52044b5228060 100755
--- a/python/paddle/distributed/fleet/base/distributed_strategy.py
+++ b/python/paddle/distributed/fleet/base/distributed_strategy.py
@@ -65,7 +65,7 @@ def check_configs_key(msg, config, field_name):
         assert key in key_list, "key:{} not in {}".format(key, field_name)
 
 
-class DistributedJobInfo(object):
+class DistributedJobInfo:
     """
     DistributedJobInfo will serialize all distributed training information
     Just for inner use: 1) debug 2) replicate experiments
@@ -106,7 +106,7 @@ ReduceStrategyFluid = paddle.fluid.BuildStrategy.ReduceStrategy
 ReduceStrategyFleet = int
 
 
-class DistributedStrategy(object):
+class DistributedStrategy:
     __lock_attr = False
 
     def __init__(self):
diff --git a/python/paddle/distributed/fleet/base/meta_optimizer_factory.py b/python/paddle/distributed/fleet/base/meta_optimizer_factory.py
index 380a9b8c177af3115b61dda74efc1089611af975..dd4611fc0a8c1daf052ed55b8361a3eadc44d559 100755
--- a/python/paddle/distributed/fleet/base/meta_optimizer_factory.py
+++ b/python/paddle/distributed/fleet/base/meta_optimizer_factory.py
@@ -26,7 +26,7 @@ meta_optimizer_names.remove("HybridParallelOptimizer")
 meta_optimizer_names.remove("HeterParallelOptimizer")
 
 
-class MetaOptimizerFactory(object):
+class MetaOptimizerFactory:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/fleet/base/role_maker.py b/python/paddle/distributed/fleet/base/role_maker.py
index 1fcd18789c0f7d38d9ccc396bf67ecbe4ad2e663..b001c5482fdfc7af804f89728f9e644db05bd233 100755
--- a/python/paddle/distributed/fleet/base/role_maker.py
+++ b/python/paddle/distributed/fleet/base/role_maker.py
@@ -35,7 +35,7 @@ class Role:
     COORDINATOR = 5
 
 
-class Gloo(object):
+class Gloo:
     """
     Gloo is a universal class for barrier and collective communication
     """
@@ -383,7 +383,7 @@ class Gloo(object):
         return output
 
 
-class RoleMakerBase(object):
+class RoleMakerBase:
     """
     RoleMakerBase is a base class for assigning a role to current process
     in distributed training.
diff --git a/python/paddle/distributed/fleet/base/runtime_factory.py b/python/paddle/distributed/fleet/base/runtime_factory.py
index e9006409674a724727098c5ef04af016e741d376..51758859035d49e1e5077f38d6987733c1cba2d9 100644
--- a/python/paddle/distributed/fleet/base/runtime_factory.py
+++ b/python/paddle/distributed/fleet/base/runtime_factory.py
@@ -17,7 +17,7 @@ from ...ps.the_one_ps import TheOnePSRuntime
 __all__ = []
 
 
-class RuntimeFactory(object):
+class RuntimeFactory:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/fleet/base/strategy_compiler.py b/python/paddle/distributed/fleet/base/strategy_compiler.py
index 348a79b9d442aa17604ef87883f8380e8b54fdd8..f1844a112398387dff06dcc340e8524584a8c913 100644
--- a/python/paddle/distributed/fleet/base/strategy_compiler.py
+++ b/python/paddle/distributed/fleet/base/strategy_compiler.py
@@ -106,7 +106,7 @@ def maximum_path_len_algo(optimizer_list):
     return candidate
 
 
-class StrategyCompilerBase(object):
+class StrategyCompilerBase:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/fleet/base/topology.py b/python/paddle/distributed/fleet/base/topology.py
index 6fa1521d58c46427d7ae8691656b5f935a4d49f5..335125123c4d6689619579a2639336cb0504a8c1 100644
--- a/python/paddle/distributed/fleet/base/topology.py
+++ b/python/paddle/distributed/fleet/base/topology.py
@@ -23,7 +23,7 @@ __all__ = ['CommunicateTopology', 'HybridCommunicateGroup']
 _HYBRID_PARALLEL_GROUP = None
 
 
-class ParallelMode(object):
+class ParallelMode:
     """
     There are all the parallel modes currently supported:
     - DATA_PARALLEL: Distribute input data to different devices.
@@ -47,7 +47,7 @@ class ParallelMode(object):
     SHARDING_PARALLEL = 3
 
 
-class CommunicateTopology(object):
+class CommunicateTopology:
     def __init__(
         self,
         hybrid_group_names=["data", "pipe", "sharding", "model"],
@@ -133,7 +133,7 @@ class CommunicateTopology(object):
         return self.get_rank(**tf)
 
 
-class HybridCommunicateGroup(object):
+class HybridCommunicateGroup:
     def __init__(self, topology):
         self.nranks = paddle.distributed.get_world_size()
         self.global_rank = paddle.distributed.get_rank()
@@ -410,7 +410,7 @@ class HybridCommunicateGroup(object):
         )
 
 
-class _CommunicateGroup(object):
+class _CommunicateGroup:
     """tmp for static"""
 
     def __init__(self):
diff --git a/python/paddle/distributed/fleet/base/util_factory.py b/python/paddle/distributed/fleet/base/util_factory.py
index dcaa256a26d11b0720f3a99a571a0bff91656c6d..8717619eafe357252fb737a3b693ce32a62ab1a7 100755
--- a/python/paddle/distributed/fleet/base/util_factory.py
+++ b/python/paddle/distributed/fleet/base/util_factory.py
@@ -31,7 +31,7 @@ import numpy as np
 __all__ = []
 
 
-class UtilFactory(object):
+class UtilFactory:
     def _create_util(self, context=None):
         util = UtilBase()
         if context is not None and "valid_strategy" in context:
@@ -41,7 +41,7 @@ class UtilFactory(object):
         return util
 
 
-class UtilBase(object):
+class UtilBase:
     def __init__(self):
         self.role_maker = None
         self.dist_strategy = None
diff --git a/python/paddle/distributed/fleet/data_generator/data_generator.py b/python/paddle/distributed/fleet/data_generator/data_generator.py
index ec6114dd21f3a5c6cab612b030866e6960616770..abf8f5f49fa1f7b911a7d377627dda0b06b2ceeb 100644
--- a/python/paddle/distributed/fleet/data_generator/data_generator.py
+++ b/python/paddle/distributed/fleet/data_generator/data_generator.py
@@ -17,7 +17,7 @@ import sys
 __all__ = []
 
 
-class DataGenerator(object):
+class DataGenerator:
     """
     DataGenerator is a general Base class for user to inherit
     A user who wants to define his/her own python processing logic
diff --git a/python/paddle/distributed/fleet/dataset/dataset.py b/python/paddle/distributed/fleet/dataset/dataset.py
index f5b3140064593adfcd50e4f38209f324eb1e1753..870c936e543a67aadcf20a4df596daf9368d2413 100755
--- a/python/paddle/distributed/fleet/dataset/dataset.py
+++ b/python/paddle/distributed/fleet/dataset/dataset.py
@@ -20,7 +20,7 @@ import paddle.fluid.core as core
 __all__ = []
 
 
-class DatasetBase(object):
+class DatasetBase:
     """Base dataset class."""
 
     def __init__(self):
diff --git a/python/paddle/distributed/fleet/dataset/index_dataset.py b/python/paddle/distributed/fleet/dataset/index_dataset.py
index 87bf2bc738dedbc5b90ed933b50b24d0b1636ac8..7df2931b5d87db69433c520a74a29af4f8b3c251 100644
--- a/python/paddle/distributed/fleet/dataset/index_dataset.py
+++ b/python/paddle/distributed/fleet/dataset/index_dataset.py
@@ -16,7 +16,7 @@ from paddle.fluid import core
 __all__ = []
 
 
-class Index(object):
+class Index:
     def __init__(self, name):
         self._name = name
 
diff --git a/python/paddle/distributed/fleet/elastic/manager.py b/python/paddle/distributed/fleet/elastic/manager.py
index b23a11642259073245148a198fe99ce63484ef2d..0344c1d43741b830300c8f889a4d8da0a7f238f0 100644
--- a/python/paddle/distributed/fleet/elastic/manager.py
+++ b/python/paddle/distributed/fleet/elastic/manager.py
@@ -52,7 +52,7 @@ class ElasticStatus:
     EXIT = "exit"
 
 
-class LauncherInterface(object):
+class LauncherInterface:
     def __init__(self, args):
         self.args = args
         self.procs = []
@@ -124,7 +124,7 @@ class LauncherInterface(object):
         raise NotImplementedError
 
 
-class ElasticManager(object):
+class ElasticManager:
     def __init__(self, args, etcd_client):
 
         self.args = args
diff --git a/python/paddle/distributed/fleet/fleet.py b/python/paddle/distributed/fleet/fleet.py
index 695f03fe1f2a06fb955f3add9cb1bfaaaf319275..617eb5729aef6451b8adb17396b3817b899f5cb6 100644
--- a/python/paddle/distributed/fleet/fleet.py
+++ b/python/paddle/distributed/fleet/fleet.py
@@ -95,7 +95,7 @@ inited_runtime_handler = wrap_decorator(_inited_runtime_handler_)
 is_non_distributed_check = wrap_decorator(_is_non_distributed_check_)
 
 
-class Fleet(object):
+class Fleet:
     """
     Unified API for distributed training of PaddlePaddle
     Please reference the https://github.com/PaddlePaddle/PaddleFleetX for details
diff --git a/python/paddle/distributed/fleet/launch_utils.py b/python/paddle/distributed/fleet/launch_utils.py
index 64795140cd7c6c52a2d1b0a85972b6b30adc6d5f..b676eee5bfb1892d66183b62f48b55d89b52c6b7 100755
--- a/python/paddle/distributed/fleet/launch_utils.py
+++ b/python/paddle/distributed/fleet/launch_utils.py
@@ -60,7 +60,7 @@ class DeviceMode:
     MLU = 4
 
 
-class Cluster(object):
+class Cluster:
     def __init__(self, hdfs):
         self.job_server = None
         self.pods = []
@@ -133,7 +133,7 @@ class Cluster(object):
         return None
 
 
-class JobServer(object):
+class JobServer:
     def __init__(self):
         self.endpoint = None
 
@@ -147,7 +147,7 @@ class JobServer(object):
         return not self == j
 
 
-class Trainer(object):
+class Trainer:
     def __init__(self):
         self.accelerators = []
         self.endpoint = None
@@ -179,7 +179,7 @@ class Trainer(object):
         return self.rank
 
 
-class Pod(object):
+class Pod:
     def __init__(self):
         self.rank = None
         self.id = None
@@ -483,7 +483,7 @@ def pretty_print_envs(envs, header=None):
     return _str
 
 
-class TrainerProc(object):
+class TrainerProc:
     def __init__(self):
         self.proc = None
         self.log_fn = None
@@ -1278,7 +1278,7 @@ def get_mapped_cluster_from_args_with_rank_mapping(args, device_mode):
     )
 
 
-class ParameterServerLauncher(object):
+class ParameterServerLauncher:
     def __init__(self, args, distribute_mode):
         self.args = args
         self.distribute_mode = distribute_mode
diff --git a/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_optimizer.py b/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_optimizer.py
index ee3f96e243fe2a1ccd88879ff25aa6017ca23a34..b0495e13b21c83a205c4e79012a9d70f245f0332 100644
--- a/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_optimizer.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_optimizer.py
@@ -24,7 +24,7 @@ HcomGroupConfig = namedtuple('HcomGroupConfig', ['name', 'nranks', 'rank_ids'])
 __all__ = []
 
 
-class AscendIRParser(object):
+class AscendIRParser:
     def __init__(self, auto_dp=False, world_rank_size=1):
         self.graph_idx = 0
         self.hcom_endpoints = {}
diff --git a/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_parser.py b/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_parser.py
index b24e51896f622f5355b2de0e54ed6f47f26d6117..6158b4a7d4108f8528f63524a84bc84362df78f2 100644
--- a/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_parser.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_parser.py
@@ -101,7 +101,7 @@ global_cnt = -1
 global_input_cnt = -1
 
 
-class AscendHelper(object):
+class AscendHelper:
     def __init__(self):
         self.dtype2ge_map = {
             0: core.GEDataType.DT_BOOL,
@@ -136,7 +136,7 @@ class AscendHelper(object):
         return self.dtype2np_map[index]
 
 
-class AscendParserFactory(object):
+class AscendParserFactory:
     def __init__(self, graph, var2geop):
         self.graph = graph
         self.var2geop = var2geop
@@ -149,7 +149,7 @@ class AscendParserFactory(object):
             raise ValueError("parser class %s does not exist" % parser_class)
 
 
-class AscendParserBase(object):
+class AscendParserBase:
     def __init__(self, graph, var2geop):
         self.graph = graph
         self.var2geop = var2geop
diff --git a/python/paddle/distributed/fleet/meta_optimizers/common.py b/python/paddle/distributed/fleet/meta_optimizers/common.py
index 7a3c89f1e9dea2b584f56f8461b2f8624d47d0cf..03ed84563b6283256ea0ac25e3f3cc861eac809b 100644
--- a/python/paddle/distributed/fleet/meta_optimizers/common.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/common.py
@@ -53,7 +53,7 @@ def is_optimizer_op(op):
     ) & int(OpRole.Optimize)
 
 
-class CollectiveHelper(object):
+class CollectiveHelper:
     def __init__(self, role_maker, nrings=1, wait_port=True):
         self.nrings = nrings
         self.wait_port = wait_port
diff --git a/python/paddle/distributed/fleet/meta_optimizers/dygraph_optimizer/dygraph_sharding_optimizer.py b/python/paddle/distributed/fleet/meta_optimizers/dygraph_optimizer/dygraph_sharding_optimizer.py
index 7e5f6983867e19a9a4cd0ff3f9313f135fe93423..63037dc6f616825cded87026bf446a419ee014e3 100755
--- a/python/paddle/distributed/fleet/meta_optimizers/dygraph_optimizer/dygraph_sharding_optimizer.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/dygraph_optimizer/dygraph_sharding_optimizer.py
@@ -25,7 +25,7 @@ def _is_trainable(param):
     return not param.stop_gradient
 
 
-class DygraphShardingOptimizer(object):
+class DygraphShardingOptimizer:
     """
     A wrapper for Sharding Optimizer in Dygraph.
 
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/fp16_helper.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/fp16_helper.py
index e4db252cf7c95132660555ea93c389fecb2e67b7..1c500ea56b5cb6242f9251960ff48f7a9d2a13df 100755
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/fp16_helper.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/fp16_helper.py
@@ -23,7 +23,7 @@ from paddle.fluid import core
 __all__ = []
 
 
-class FP16Utils(object):
+class FP16Utils:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/gradient_clip_helper.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/gradient_clip_helper.py
index 563757d35f4057a5b2f1b444fe09677388ebda56..288e9d7d8a4d96065a487f06e88d5f45ed908e60 100755
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/gradient_clip_helper.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/gradient_clip_helper.py
@@ -17,7 +17,7 @@ from paddle.distributed.fleet.meta_optimizers.common import OP_ROLE_KEY, OpRole
 __all__ = []
 
 
-class GradientClipHelper(object):
+class GradientClipHelper:
     def __init__(self, mp_ring_id):
         self.mp_ring_id = mp_ring_id
 
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/offload_helper.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/offload_helper.py
index ac10bb42383d1598c51485972bc1e723f907ed48..c1951299c2cdc4311f676aba3912cb673cb74946 100755
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/offload_helper.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/offload_helper.py
@@ -44,7 +44,7 @@ class PlaceType:
         return PlaceType.CPU
 
 
-class OffloadHelper(object):
+class OffloadHelper:
     cpu_place_type = 0
     cuda_place_type = PlaceType.default_device()
     cuda_pinned_place_type = PlaceType.default_pinned()
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/prune.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/prune.py
index 895fd2f7acb2712afef26b12cb07661a18869a3d..9a264a7dd1ba0c419218513c80110380853bcb7d 100755
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/prune.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/prune.py
@@ -15,7 +15,7 @@
 __all__ = []
 
 
-class ProgramDeps(object):
+class ProgramDeps:
     def __init__(self, block, start_vars, end_vars):
         self._block = block
         # vars where to start to build the deps
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/shard.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/shard.py
index d33d04098d051a73aff63bcd268d14382d761bc7..82a7a7494d5e66476726bfdbcfdade2f73c2c3e7 100644
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/shard.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/shard.py
@@ -22,7 +22,7 @@ from paddle.distributed.fleet.meta_optimizers.sharding.fp16_helper import (
 __all__ = []
 
 
-class Shard(object):
+class Shard:
     def __init__(
         self,
     ):
@@ -155,7 +155,7 @@ class Shard(object):
         return grads_in_shard
 
 
-class ProgramSegment(object):
+class ProgramSegment:
     def __init__(self, block):
         self._block = block
         self._allreduce_vars = []
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/utils.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/utils.py
index ea42130300f11208f0f75917e3615b9d9fb86dd7..9feed7b1e5aad83f6ae67319347d791b7422bd3e 100755
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/utils.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/utils.py
@@ -408,7 +408,7 @@ def insert_allreduce_ops(
     return
 
 
-class FuseHelper(object):
+class FuseHelper:
     @staticmethod
     def sort_vars_by_dtype(block, vars_name):
         fp32_vars = []
diff --git a/python/paddle/distributed/fleet/meta_optimizers/sharding/weight_decay_helper.py b/python/paddle/distributed/fleet/meta_optimizers/sharding/weight_decay_helper.py
index 3d5d8aa2a3851bc1ece59ae8443a5f260d91e3b7..0a841cf243d14bf7a1cdb7135938699e14074c05 100644
--- a/python/paddle/distributed/fleet/meta_optimizers/sharding/weight_decay_helper.py
+++ b/python/paddle/distributed/fleet/meta_optimizers/sharding/weight_decay_helper.py
@@ -17,7 +17,7 @@ from paddle.distributed.fleet.meta_optimizers.common import OP_ROLE_VAR_KEY
 __all__ = []
 
 
-class WeightDecayHelper(object):
+class WeightDecayHelper:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/fleet/meta_parallel/parallel_layers/pp_layers.py b/python/paddle/distributed/fleet/meta_parallel/parallel_layers/pp_layers.py
index 663cd7d28140abe77b0a9c61ee515ecc58131948..7ddbb64883d9144f9fdf6a8658220c54ccd42ff3 100755
--- a/python/paddle/distributed/fleet/meta_parallel/parallel_layers/pp_layers.py
+++ b/python/paddle/distributed/fleet/meta_parallel/parallel_layers/pp_layers.py
@@ -53,7 +53,7 @@ from paddle.incubate.distributed.fleet import recompute_hybrid
 __all__ = []
 
 
-class LayerDesc(object):
+class LayerDesc:
     def __init__(self, layer_func, *inputs, **kwargs):
         self.layer_func = layer_func
         self.inputs = inputs
@@ -89,7 +89,7 @@ class SharedLayerDesc(LayerDesc):
         self.shared_weight_attr = shared_weight_attr
 
 
-class SegmentLayers(object):
+class SegmentLayers:
     def __init__(
         self,
         layers_desc,
diff --git a/python/paddle/distributed/fleet/runtime/runtime_base.py b/python/paddle/distributed/fleet/runtime/runtime_base.py
index 2e8bacfbc3b1ded58e63e8d9e93764a0c0090b91..192754d6638b9d09a1882c3650eab22d7ca2a9ef 100644
--- a/python/paddle/distributed/fleet/runtime/runtime_base.py
+++ b/python/paddle/distributed/fleet/runtime/runtime_base.py
@@ -15,7 +15,7 @@
 __all__ = []
 
 
-class RuntimeBase(object):
+class RuntimeBase:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/fleet/utils/fs.py b/python/paddle/distributed/fleet/utils/fs.py
index 8a67301e174618832215e564d553a771b50d3a2e..667752e668a96f753a81013128344b77e101cbb2 100644
--- a/python/paddle/distributed/fleet/utils/fs.py
+++ b/python/paddle/distributed/fleet/utils/fs.py
@@ -46,7 +46,7 @@ class FSShellCmdAborted(ExecuteError):
     pass
 
 
-class FS(object):
+class FS:
     @abc.abstractmethod
     def ls_dir(self, fs_path):
         raise NotImplementedError
diff --git a/python/paddle/distributed/fleet/utils/http_server.py b/python/paddle/distributed/fleet/utils/http_server.py
index 5602c5f01afc86137c337ba9b7c4cc895191dde5..2828b9e5ddfecf9b92540f14281431b3ef258e12 100644
--- a/python/paddle/distributed/fleet/utils/http_server.py
+++ b/python/paddle/distributed/fleet/utils/http_server.py
@@ -128,7 +128,7 @@ class KVHandler(SimpleHTTPServer.SimpleHTTPRequestHandler):
         self.end_headers()
 
 
-class KVHTTPServer(HTTPServer, object):
+class KVHTTPServer(HTTPServer):
     """
     it is a http server storing kv pairs.
     """
diff --git a/python/paddle/distributed/fleet/utils/hybrid_parallel_inference.py b/python/paddle/distributed/fleet/utils/hybrid_parallel_inference.py
index a56c71fc40b5d0ec6855863061ef7c56c82adee4..5ba7c9c29762adfb5a83be356ff9817e77852011 100644
--- a/python/paddle/distributed/fleet/utils/hybrid_parallel_inference.py
+++ b/python/paddle/distributed/fleet/utils/hybrid_parallel_inference.py
@@ -20,7 +20,7 @@ import paddle.distributed.fleet as fleet
 import numpy as np
 
 
-class HybridParallelInferenceHelper(object):
+class HybridParallelInferenceHelper:
     """
     A helper class to split program for inference with hybrid parallelism.
 
diff --git a/python/paddle/distributed/launch/context/__init__.py b/python/paddle/distributed/launch/context/__init__.py
index 037fc0efbc51be8c732b757d9570b869f0c11125..0d3410e368e7093cc57714be69483a111f3cb6d5 100644
--- a/python/paddle/distributed/launch/context/__init__.py
+++ b/python/paddle/distributed/launch/context/__init__.py
@@ -21,7 +21,7 @@ from .args_envs import parse_args, fetch_envs, env_args_mapping
 import logging
 
 
-class Context(object):
+class Context:
     def __init__(self, enable_plugin=True):
         self.args, self.unknown_args = parse_args()
         self.envs = fetch_envs()
diff --git a/python/paddle/distributed/launch/context/device.py b/python/paddle/distributed/launch/context/device.py
index f05bc1b776869675915464da95ad483798c89543..2708755c2eb7a29aa26c0ab0d7d7e0d799e41546 100644
--- a/python/paddle/distributed/launch/context/device.py
+++ b/python/paddle/distributed/launch/context/device.py
@@ -27,7 +27,7 @@ class DeviceType:
     CUSTOM_DEVICE = 'custom_device'
 
 
-class Device(object):
+class Device:
     def __init__(self, dtype=None, memory="", labels=""):
         self._dtype = dtype
         self._memory = memory
diff --git a/python/paddle/distributed/launch/context/event.py b/python/paddle/distributed/launch/context/event.py
index 3859edb83f29c1320ff90a91253d56d1f0a04a2a..7792907e47b0af9ebbd70804c6ce2276800aff7c 100644
--- a/python/paddle/distributed/launch/context/event.py
+++ b/python/paddle/distributed/launch/context/event.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class Event(object):
+class Event:
     def __init__(self, kind="status", message="", fatal=False):
         self.kind = kind
         self.message = message
diff --git a/python/paddle/distributed/launch/context/node.py b/python/paddle/distributed/launch/context/node.py
index b8c1a2a14f5ce64e1eb05464012404d75a50b8d3..04748d2935bcc968be50dec9378a841d75803657 100644
--- a/python/paddle/distributed/launch/context/node.py
+++ b/python/paddle/distributed/launch/context/node.py
@@ -21,7 +21,7 @@ import struct
 from contextlib import closing
 
 
-class Node(object):
+class Node:
     def __init__(self):
         # self.device = Device.detect_device()
         self.device = Device.parse_device()
diff --git a/python/paddle/distributed/launch/context/resource.py b/python/paddle/distributed/launch/context/resource.py
index b7baf6fd7b621c647c6708843c10fc24a32e1b5e..3b4d16e7b592047bb6f606b0f6cedffcdb013717 100644
--- a/python/paddle/distributed/launch/context/resource.py
+++ b/python/paddle/distributed/launch/context/resource.py
@@ -13,6 +13,6 @@
 # limitations under the License.
 
 
-class Resource(object):
+class Resource:
     def __init__(self):
         self.devices = []
diff --git a/python/paddle/distributed/launch/context/status.py b/python/paddle/distributed/launch/context/status.py
index b87b7b3fb82d8ea76e63d27ad83a7edbe4a611b0..460f14f0cbfe1419419548eb0d4cd3bde9b719af 100644
--- a/python/paddle/distributed/launch/context/status.py
+++ b/python/paddle/distributed/launch/context/status.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class Status(object):
+class Status:
     UNINIT = "uninit"
     READY = "ready"
     RUNNING = "running"
diff --git a/python/paddle/distributed/launch/controllers/controller.py b/python/paddle/distributed/launch/controllers/controller.py
index 9ff18d5b5d511f88ea45a443d0a8f070ea61caea..73954adaab741351d4df2ff22c7732a66fc010b4 100644
--- a/python/paddle/distributed/launch/controllers/controller.py
+++ b/python/paddle/distributed/launch/controllers/controller.py
@@ -31,7 +31,7 @@ class ControleMode:
     RPC = "rpc"
 
 
-class ControllerBase(object):
+class ControllerBase:
     def __init__(self, ctx):
         signal.signal(signal.SIGTERM, self.signal_handler)
         signal.signal(signal.SIGABRT, self.signal_handler)
diff --git a/python/paddle/distributed/launch/controllers/master.py b/python/paddle/distributed/launch/controllers/master.py
index 9c3f0a8501a3ac2334452ad5152bb58986ecfa33..37c6f7a50872be2a442e2a2fcb7ce34d50d6d3de 100644
--- a/python/paddle/distributed/launch/controllers/master.py
+++ b/python/paddle/distributed/launch/controllers/master.py
@@ -24,7 +24,7 @@ import random
 ETCD_PROTOCAL = 'etcd://'
 
 
-class Master(object):
+class Master:
     '''
     Master is a distributed store design to exchange info among nodes
     '''
diff --git a/python/paddle/distributed/launch/controllers/watcher.py b/python/paddle/distributed/launch/controllers/watcher.py
index a9c1f509666875bcac3f43fb455f6d97b49af6fa..c76a428d7489a7dc6358db6ab67a2195b47e7b97 100644
--- a/python/paddle/distributed/launch/controllers/watcher.py
+++ b/python/paddle/distributed/launch/controllers/watcher.py
@@ -19,7 +19,7 @@ import os
 from threading import Thread
 
 
-class Watcher(object):
+class Watcher:
     def __init__(self, ctx):
         self.ctx = ctx
 
diff --git a/python/paddle/distributed/launch/job/container.py b/python/paddle/distributed/launch/job/container.py
index 6eb313ea579657abed1dff9d79df433b8d8c3247..c78c3323a8597e0fe0a9fa484b76442672aa045f 100644
--- a/python/paddle/distributed/launch/job/container.py
+++ b/python/paddle/distributed/launch/job/container.py
@@ -20,7 +20,7 @@ import os
 import sys
 
 
-class Container(object):
+class Container:
     '''
     TODO(kuizhiqing) A container can be run by process/thread or just a callable function
     '''
diff --git a/python/paddle/distributed/launch/job/job.py b/python/paddle/distributed/launch/job/job.py
index f5c805e31bf63d00d24b459b18ccc952df5970e3..261e6ee7f292c7277b5dc3dc7715fb6552a27820 100644
--- a/python/paddle/distributed/launch/job/job.py
+++ b/python/paddle/distributed/launch/job/job.py
@@ -19,7 +19,7 @@ class JobMode:
     HETER = 'heter'
 
 
-class Job(object):
+class Job:
     def __init__(self, jid='default', mode=JobMode.COLLECTIVE, nnodes="1"):
         self._mode = mode
         self._id = jid
diff --git a/python/paddle/distributed/launch/job/pod.py b/python/paddle/distributed/launch/job/pod.py
index b65aad6e0fcb43a869d5038489db927e13261216..a322bcdccfe6d7428d1719daebde031e3850183f 100644
--- a/python/paddle/distributed/launch/job/pod.py
+++ b/python/paddle/distributed/launch/job/pod.py
@@ -20,7 +20,7 @@ import random
 import time
 
 
-class PodSepc(object):
+class PodSepc:
     def __init__(self):
         self._name = ''.join(
             random.choice('abcdefghijklmnopqrstuvwxyz') for _ in range(6)
diff --git a/python/paddle/distributed/launch/job/status.py b/python/paddle/distributed/launch/job/status.py
index 88fd09bbf2267a8f6aa69857d982ba25d787259b..d8a346ea292a4601daf3fab4d4ec402b9864c643 100644
--- a/python/paddle/distributed/launch/job/status.py
+++ b/python/paddle/distributed/launch/job/status.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class Status(object):
+class Status:
     UNINIT = "uninit"
     READY = "ready"
     RUNNING = "running"
diff --git a/python/paddle/distributed/launch/utils/kv_client.py b/python/paddle/distributed/launch/utils/kv_client.py
index 8ed46053de0d32eaa6b432254b455594b133d850..b60970382aee807bb5a26dab1f762e13bd9c0a53 100644
--- a/python/paddle/distributed/launch/utils/kv_client.py
+++ b/python/paddle/distributed/launch/utils/kv_client.py
@@ -16,7 +16,7 @@ import requests
 import time
 
 
-class KVClient(object):
+class KVClient:
     def __init__(self, endpoint='localhost:2379'):
         self.endpoint = (
             endpoint
diff --git a/python/paddle/distributed/launch/utils/kv_server.py b/python/paddle/distributed/launch/utils/kv_server.py
index d27836f1bfe6f327e50723e1a0aec073aab1c905..90dcbcf9375230efca0d7b8bd6f60af5792bf2fb 100644
--- a/python/paddle/distributed/launch/utils/kv_server.py
+++ b/python/paddle/distributed/launch/utils/kv_server.py
@@ -67,7 +67,7 @@ class KVHandler(SimpleHTTPServer.SimpleHTTPRequestHandler):
         return
 
 
-class KVServer(HTTPServer, object):
+class KVServer(HTTPServer):
     def __init__(self, port):
         super().__init__(('', port), KVHandler)
         self.kv_lock = threading.Lock()
diff --git a/python/paddle/distributed/launch/utils/nvsmi.py b/python/paddle/distributed/launch/utils/nvsmi.py
index d1a14f11666d2521438d84b8132c830e39ef1270..762870baa15a1272fe036273020b4c1cfb78c1dc 100644
--- a/python/paddle/distributed/launch/utils/nvsmi.py
+++ b/python/paddle/distributed/launch/utils/nvsmi.py
@@ -18,7 +18,7 @@ import json
 import shutil
 
 
-class Info(object):
+class Info:
     def __repr__(self):
         return str(self.__dict__)
 
diff --git a/python/paddle/distributed/launch/utils/process_context.py b/python/paddle/distributed/launch/utils/process_context.py
index 682a857f2ee6c3bfbe2e1f89f68fb13da4598852..3a8c52851778c5917fe9f88ab3d7dd004bbb1841 100644
--- a/python/paddle/distributed/launch/utils/process_context.py
+++ b/python/paddle/distributed/launch/utils/process_context.py
@@ -16,7 +16,7 @@ import subprocess
 import os, sys, signal, time
 
 
-class ProcessContext(object):
+class ProcessContext:
     def __init__(
         self,
         cmd,
diff --git a/python/paddle/distributed/passes/auto_parallel_amp.py b/python/paddle/distributed/passes/auto_parallel_amp.py
index 85f8ec2e1fd96134c2f5cd7cec8ee16a5c9d2e57..d15d38abee9133049ba63634575bc286391fde6e 100644
--- a/python/paddle/distributed/passes/auto_parallel_amp.py
+++ b/python/paddle/distributed/passes/auto_parallel_amp.py
@@ -54,7 +54,7 @@ from ..auto_parallel.utils import is_forward_op, is_backward_op, is_loss_op
 world_process_group = get_world_process_group()
 
 
-class AMPState(object):
+class AMPState:
     def __init__(self, block):
         self._block = block
         self._op_fp16_dict = (
diff --git a/python/paddle/distributed/passes/auto_parallel_data_parallel_optimization.py b/python/paddle/distributed/passes/auto_parallel_data_parallel_optimization.py
index c2779d3557199ade3749278a918ea15b34e1b389..ec3d799ee84484aa65b7d162986f059be784ccd3 100644
--- a/python/paddle/distributed/passes/auto_parallel_data_parallel_optimization.py
+++ b/python/paddle/distributed/passes/auto_parallel_data_parallel_optimization.py
@@ -572,7 +572,7 @@ class DataParallelOptimizationPass(PassBase):
             self._logger.info("individual gradient {}".format(individual_grads))
 
 
-class GradientsGroup(object):
+class GradientsGroup:
     def __init__(self, ops, max_group_size):
         self.max_group_size = max_group_size
         self.ops = ops
diff --git a/python/paddle/distributed/passes/auto_parallel_fp16.py b/python/paddle/distributed/passes/auto_parallel_fp16.py
index cdb5f4221237f720809fe200e38fd1d7d568109a..a952986c21dca717a9ac12e099fe276db06df771 100644
--- a/python/paddle/distributed/passes/auto_parallel_fp16.py
+++ b/python/paddle/distributed/passes/auto_parallel_fp16.py
@@ -126,7 +126,7 @@ def _keep_fp32_output(op, out_name):
     return False
 
 
-class FP16State(object):
+class FP16State:
     def __init__(
         self,
         program,
diff --git a/python/paddle/distributed/passes/auto_parallel_grad_clip.py b/python/paddle/distributed/passes/auto_parallel_grad_clip.py
index d570bf9c3fc1b53a74d953ad18fd2e3787fb7e8a..73432baa1d3c3d17d7aa8f9f22b3b9a9b3cb5309 100644
--- a/python/paddle/distributed/passes/auto_parallel_grad_clip.py
+++ b/python/paddle/distributed/passes/auto_parallel_grad_clip.py
@@ -143,7 +143,7 @@ def _is_about_global_norm(
     return rank_id in complete_param_ranks
 
 
-class ClipHelper(object):
+class ClipHelper:
     def __init__(self, params_grads, rank_id, block, dist_context):
         params, _ = zip(*params_grads)
         self.params = list(params)
diff --git a/python/paddle/distributed/passes/auto_parallel_sharding.py b/python/paddle/distributed/passes/auto_parallel_sharding.py
index 290a1741caed32c6811285c1c7aab55bbc7354c5..a80af73c2bced73f9de8ece59e041cafbe6db4fc 100644
--- a/python/paddle/distributed/passes/auto_parallel_sharding.py
+++ b/python/paddle/distributed/passes/auto_parallel_sharding.py
@@ -850,7 +850,7 @@ def shard_parameters(params, group_size):
     return mapping
 
 
-class ShardingInfo(object):
+class ShardingInfo:
     def __init__(self, group, rank, params_grads):
         self.group = group
         self.params_grads = dict([(p.name, (p, g)) for p, g in params_grads])
diff --git a/python/paddle/distributed/ps/coordinator.py b/python/paddle/distributed/ps/coordinator.py
index d73193845e95f90273ea124038e9418d038060b8..612502edad9d34b96140e922db1f33b2619c3240 100755
--- a/python/paddle/distributed/ps/coordinator.py
+++ b/python/paddle/distributed/ps/coordinator.py
@@ -353,7 +353,7 @@ class FLClient(FLClientBase):
             f.write(str(self.train_statical_info))
 
 
-class Coordinator(object):
+class Coordinator:
     def __init__(self, ps_hosts):
         self._communicator = FLCommunicator(ps_hosts)
         self._client_selector = None
diff --git a/python/paddle/distributed/ps/the_one_ps.py b/python/paddle/distributed/ps/the_one_ps.py
index b4881959ee5733ff3913ee6c16e3c37bea3a9178..3dbbae864789a739433a5a349ca1067c2dada90c 100755
--- a/python/paddle/distributed/ps/the_one_ps.py
+++ b/python/paddle/distributed/ps/the_one_ps.py
@@ -914,7 +914,7 @@ class fsClient:
         proto.hadoop_bin = self.fs_client_param.hadoop_bin
 
 
-class PsDescBuilder(object):
+class PsDescBuilder:
     def __init__(self, context):
         self.context = context
         self.is_sync = context['is_sync']
diff --git a/python/paddle/distributed/ps/utils/ps_factory.py b/python/paddle/distributed/ps/utils/ps_factory.py
index 0726fe15dc4cd437810b22b64ae5cfdb6dd04540..0161352ec95dc9bc912b961c76d46c4b32ddf92e 100755
--- a/python/paddle/distributed/ps/utils/ps_factory.py
+++ b/python/paddle/distributed/ps/utils/ps_factory.py
@@ -27,7 +27,7 @@ __all__ = [
 ]
 
 
-class PsProgramBuilderFactory(object):
+class PsProgramBuilderFactory:
     def __init__(self):
         pass
 
diff --git a/python/paddle/distributed/ps/utils/ps_program_builder.py b/python/paddle/distributed/ps/utils/ps_program_builder.py
index a07e2ebe1408a7515d1fdded23aebedf77fa8559..1831f7061b6e0c2badbdec2b4ec0585862f7d644 100755
--- a/python/paddle/distributed/ps/utils/ps_program_builder.py
+++ b/python/paddle/distributed/ps/utils/ps_program_builder.py
@@ -19,7 +19,7 @@ from paddle.distributed.fleet.base.private_helper_function import (
 from paddle.distributed.passes import new_pass
 
 
-class PsProgramBuilder(object):
+class PsProgramBuilder:
     def __init__(self, pass_ctx):
         self.pass_ctx = pass_ctx
         self.attrs = self.pass_ctx._attrs
diff --git a/python/paddle/distributed/ps/utils/public.py b/python/paddle/distributed/ps/utils/public.py
index 578d664dc4137bdc79da8af539178a3ae20503ed..3b3a44d49890eb02ccb80fe624e64d99a0fc0a73 100755
--- a/python/paddle/distributed/ps/utils/public.py
+++ b/python/paddle/distributed/ps/utils/public.py
@@ -88,7 +88,7 @@ class DistributedMode:
     NU = 5
 
 
-class TrainerRuntimeConfig(object):
+class TrainerRuntimeConfig:
     def __init__(self, valid_strategy):
         self.mode = None
         num_threads = os.getenv("CPU_NUM", "1")
diff --git a/python/paddle/distributed/spawn.py b/python/paddle/distributed/spawn.py
index 2df9118ac4e870a7ff2ae5d24c099eeaf4c4bc83..a371f5d559b597ea06171eefc6fc69dcbe56e6bf 100644
--- a/python/paddle/distributed/spawn.py
+++ b/python/paddle/distributed/spawn.py
@@ -43,7 +43,7 @@ from paddle.fluid.framework import set_flags
 __all__ = []
 
 
-class ParallelEnvArgs(object):
+class ParallelEnvArgs:
     def __init__(self):
         # Paddle cluster nodes ips, such as 192.168.0.16,192.168.0.17..
         self.cluster_node_ips = None
@@ -412,7 +412,7 @@ def _func_wrapper(func, args, error_queue, return_queue, env_dict, backend):
         sys.exit(1)
 
 
-class MultiprocessContext(object):
+class MultiprocessContext:
     def __init__(self, processes, error_queues, return_queues):
         _py_supported_check()
         self.error_queues = error_queues
diff --git a/python/paddle/distributed/utils/launch_utils.py b/python/paddle/distributed/utils/launch_utils.py
index 975f5d4935644d0fe3f3abc7c10519517b2f234f..0c1ab76e5506d1b3896bb57f8d731acd3787d390 100644
--- a/python/paddle/distributed/utils/launch_utils.py
+++ b/python/paddle/distributed/utils/launch_utils.py
@@ -99,7 +99,7 @@ def get_gpus(selected_gpus):
     return gpus
 
 
-class Hdfs(object):
+class Hdfs:
     def __init__(self):
         self.hdfs_ugi = None
         self.hdfs_name = None
@@ -128,7 +128,7 @@ class Hdfs(object):
         return not self == n
 
 
-class Cluster(object):
+class Cluster:
     def __init__(self, hdfs):
         self.job_server = None
         self.pods = []
@@ -194,7 +194,7 @@ class Cluster(object):
         return None
 
 
-class JobServer(object):
+class JobServer:
     def __init__(self):
         self.endpoint = None
 
@@ -208,7 +208,7 @@ class JobServer(object):
         return not self == j
 
 
-class Trainer(object):
+class Trainer:
     def __init__(self):
         self.gpus = []
         self.endpoint = None
@@ -239,7 +239,7 @@ class Trainer(object):
         return self.rank
 
 
-class Pod(object):
+class Pod:
     def __init__(self):
         self.rank = None
         self.id = None
@@ -454,7 +454,7 @@ def _prepare_trainer_env(cluster, trainer, backend=None):
     return proc_env
 
 
-class TrainerProc(object):
+class TrainerProc:
     def __init__(self):
         self.proc = None
         self.log_fn = None
diff --git a/python/paddle/distribution/constraint.py b/python/paddle/distribution/constraint.py
index 46f919f3858592ce9152dc1fa426bc863a9214c2..092faf16937b4e2a25ccc9b309d64a302932537f 100644
--- a/python/paddle/distribution/constraint.py
+++ b/python/paddle/distribution/constraint.py
@@ -14,7 +14,7 @@
 import paddle
 
 
-class Constraint(object):
+class Constraint:
     """Constraint condition for random variable."""
 
     def __call__(self, value):
diff --git a/python/paddle/distribution/distribution.py b/python/paddle/distribution/distribution.py
index 92dd306ce022f68dc95c09a8b21ed2297ab47482..27febe07a3810c1aec990c6a0b8abe20a0e68775 100644
--- a/python/paddle/distribution/distribution.py
+++ b/python/paddle/distribution/distribution.py
@@ -33,7 +33,7 @@ from paddle.fluid.framework import (
 from paddle.fluid.layers import tensor
 
 
-class Distribution(object):
+class Distribution:
     """
     The abstract base class for probability distributions. Functions are
     implemented in specific distributions.
diff --git a/python/paddle/distribution/kl.py b/python/paddle/distribution/kl.py
index 4f95366391a155eb3365002529e594bf9fed3492..cf8857629893becfaee8cb69136cdd8142fb3132 100644
--- a/python/paddle/distribution/kl.py
+++ b/python/paddle/distribution/kl.py
@@ -127,7 +127,7 @@ def _dispatch(cls_p, cls_q):
 
 
 @functools.total_ordering
-class _Compare(object):
+class _Compare:
     def __init__(self, *classes):
         self.classes = classes
 
diff --git a/python/paddle/distribution/transform.py b/python/paddle/distribution/transform.py
index f13c224691593a132a0f97e07640b84bc626145c..db55eca2d1a71c837ec3e94bf4eb4145d693c96e 100644
--- a/python/paddle/distribution/transform.py
+++ b/python/paddle/distribution/transform.py
@@ -58,7 +58,7 @@ class Type(enum.Enum):
         return _type in (cls.BIJECTION, cls.INJECTION)
 
 
-class Transform(object):
+class Transform:
     r"""Base class for the transformations of random variables.
 
     ``Transform`` can be used to represent any differentiable and injective
diff --git a/python/paddle/distribution/variable.py b/python/paddle/distribution/variable.py
index e7aa1e1a680a6a58bf73a9e3bed605119743cd42..99cafc5ea788ec29a3d09ba0d7ffba8bc89fe2e7 100644
--- a/python/paddle/distribution/variable.py
+++ b/python/paddle/distribution/variable.py
@@ -15,7 +15,7 @@
 from paddle.distribution import constraint
 
 
-class Variable(object):
+class Variable:
     """Random variable of probability distribution.
 
     Args:
diff --git a/python/paddle/fluid/average.py b/python/paddle/fluid/average.py
index 2d83246df7f636036a1e6d886bb349c130da291e..32db4ba0a9093b5376861e016c891acd545bcc44 100644
--- a/python/paddle/fluid/average.py
+++ b/python/paddle/fluid/average.py
@@ -39,7 +39,7 @@ def _is_number_or_matrix_(var):
     return _is_number_(var) or isinstance(var, np.ndarray)
 
 
-class WeightedAverage(object):
+class WeightedAverage:
     """
     Calculate weighted average.
 
diff --git a/python/paddle/fluid/backward.py b/python/paddle/fluid/backward.py
index e73e2fe1ab10b1cef7d35541c7e0eb0203d97012..279ac480453dd7b2ced706e5936fb438198d5f66 100755
--- a/python/paddle/fluid/backward.py
+++ b/python/paddle/fluid/backward.py
@@ -41,7 +41,7 @@ _logger = log_helper.get_logger(
 )
 
 
-class ProgramStats(object):
+class ProgramStats:
     def __init__(self, block, ops):
         self.block = block
         self.ops = ops
@@ -789,7 +789,7 @@ def _find_not_need_ops(grad_op_descs, forward_ops, input_grad_names_set):
         (set[core.OpDesc]): A set of OpDescs which should be pruned.
     """
 
-    class Var(object):
+    class Var:
         def __init__(self, var_name):
             self.var_name = var_name
             self.gen_op = None
@@ -804,7 +804,7 @@ def _find_not_need_ops(grad_op_descs, forward_ops, input_grad_names_set):
             assert isinstance(op, Op)
             self.pendding_ops.append(op)
 
-    class Op(object):
+    class Op:
         def __init__(self, op_desc):
             self.op_desc = op_desc
             self.inputs = []
diff --git a/python/paddle/fluid/clip.py b/python/paddle/fluid/clip.py
index 1ad98da957213735b6ebea3eab2334745fd5951e..68a2f8a0deea36274a1ab2db2796cdcb122fe706 100644
--- a/python/paddle/fluid/clip.py
+++ b/python/paddle/fluid/clip.py
@@ -92,7 +92,7 @@ def _squared_l2_norm(x):
     return out
 
 
-class BaseErrorClipAttr(object):
+class BaseErrorClipAttr:
     def __str__(self):
         raise NotImplementedError()
 
@@ -177,7 +177,7 @@ def error_clip_callback(block, context):
             error_clip._append_clip_op(block, grad_n)
 
 
-class ClipGradBase(object):
+class ClipGradBase:
     def __init__(self):
         super().__init__()
 
diff --git a/python/paddle/fluid/communicator.py b/python/paddle/fluid/communicator.py
index bd43dce83711188deb44e158c9fa9bcf3633911e..55733c87f67c9b557bc156fefcb08aed27b8975e 100755
--- a/python/paddle/fluid/communicator.py
+++ b/python/paddle/fluid/communicator.py
@@ -38,7 +38,7 @@ from paddle.fluid.incubate.fleet.parameter_server.mode import DistributedMode
 __all__ = ['Communicator', 'FLCommunicator', 'LargeScaleKV']
 
 
-class Communicator(object):
+class Communicator:
     def __init__(self, mode, kwargs=None, envs=None):
         """
         Communicator is used for async distribute training in distribute_transpiler mode.
@@ -246,7 +246,7 @@ class FLCommunicator(Communicator):  ## only for coordinator
         return info_mp
 
 
-class LargeScaleKV(object):
+class LargeScaleKV:
     def __init__(self):
         self.scale_kv = core.LargeScaleKV()
 
@@ -260,7 +260,7 @@ class LargeScaleKV(object):
         return self.scale_kv.size(varname)
 
 
-class HeterClient(object):
+class HeterClient:
     def __init__(self, endpoint, previous_endpoint, trainer_id):
         self.heter_client_ = core.HeterClient(
             endpoint, previous_endpoint, trainer_id
diff --git a/python/paddle/fluid/compiler.py b/python/paddle/fluid/compiler.py
index 1adef41f86a35065bf6f44b48d50875ff9aa71ed..47110b1e0bf7e58d503011be28e6cbcc9ae3a673 100644
--- a/python/paddle/fluid/compiler.py
+++ b/python/paddle/fluid/compiler.py
@@ -105,7 +105,7 @@ def _should_broadcast_or_not_exists(program, var_name):
     return not is_distributed
 
 
-class CompiledProgram(object):
+class CompiledProgram:
     """
     :api_attr: Static Graph
 
@@ -567,7 +567,7 @@ class CompiledProgram(object):
         return place_list
 
 
-class IpuDynamicPatcher(object):
+class IpuDynamicPatcher:
     """
     Patcher for IPU dynamic2static support.
     """
@@ -777,7 +777,7 @@ class IpuDynamicPatcher(object):
             setattr(module, key, attr)
 
 
-class IpuStrategy(object):
+class IpuStrategy:
     """
     Help users precisely control the graph building in :code:`paddle.static.IpuCompiledProgram` .
 
@@ -1237,7 +1237,7 @@ class IpuStrategy(object):
         return self.get_option('enable_fp16')
 
 
-class IpuCompiledProgram(object):
+class IpuCompiledProgram:
     """
     The IpuCompiledProgram is used to transform a program to a ipu-target program,
     such as forward graph extraction, computing graph transformation, useless scale Ops clean, etc.
diff --git a/python/paddle/fluid/contrib/decoder/beam_search_decoder.py b/python/paddle/fluid/contrib/decoder/beam_search_decoder.py
index 429feda47bb53d9c750d1ff7de69ff4ed7fbba6f..717d31c2fe1b9766db0ab90544e19abd531e586b 100644
--- a/python/paddle/fluid/contrib/decoder/beam_search_decoder.py
+++ b/python/paddle/fluid/contrib/decoder/beam_search_decoder.py
@@ -37,7 +37,7 @@ class _DecoderType:
     BEAM_SEARCH = 2
 
 
-class InitState(object):
+class InitState:
     """
     The initial hidden state object. The state objects holds a variable, and may
     use it to initialize the hidden state cell of RNN. Usually used as input to
@@ -98,7 +98,7 @@ class InitState(object):
         return self._need_reorder
 
 
-class _MemoryState(object):
+class _MemoryState:
     def __init__(self, state_name, rnn_obj, init_state):
         self._state_name = state_name  # each is a rnn.memory
         self._rnn_obj = rnn_obj
@@ -113,7 +113,7 @@ class _MemoryState(object):
         self._rnn_obj.update_memory(self._state_mem, state)
 
 
-class _ArrayState(object):
+class _ArrayState:
     def __init__(self, state_name, block, init_state):
         self._state_name = state_name
         self._block = block
@@ -161,7 +161,7 @@ class _ArrayState(object):
         layers.array_write(state, array=self._state_array, i=self._counter)
 
 
-class StateCell(object):
+class StateCell:
     """
     The state cell class stores the hidden state of the RNN cell. A typical RNN
     cell has one or more hidden states, and one or more step inputs. This class
@@ -401,7 +401,7 @@ class StateCell(object):
         return self._cur_states[self._out_state]
 
 
-class TrainingDecoder(object):
+class TrainingDecoder:
     """
     A decoder that can only be used for training. The decoder could be
     initialized with a `StateCell` object. The computation within the RNN cell
@@ -547,7 +547,7 @@ class TrainingDecoder(object):
             )
 
 
-class BeamSearchDecoder(object):
+class BeamSearchDecoder:
     """
     A beam search decoder that can be used for inference. The decoder should be
     initialized with a `StateCell` object. The decode process can be defined
diff --git a/python/paddle/fluid/contrib/extend_optimizer/extend_optimizer_with_weight_decay.py b/python/paddle/fluid/contrib/extend_optimizer/extend_optimizer_with_weight_decay.py
index e85f3e45a24108528d0e622addd7a8c5e32664b5..53a010c23ce9ddb61de407d2b385868e034e71fd 100644
--- a/python/paddle/fluid/contrib/extend_optimizer/extend_optimizer_with_weight_decay.py
+++ b/python/paddle/fluid/contrib/extend_optimizer/extend_optimizer_with_weight_decay.py
@@ -17,7 +17,7 @@ from paddle.fluid import framework as framework
 __all__ = ["extend_with_decoupled_weight_decay"]
 
 
-class DecoupledWeightDecay(object):
+class DecoupledWeightDecay:
     def __init__(self, coeff=0.0, apply_decay_param_fun=None, **kwargs):
         if not isinstance(coeff, float) and not isinstance(
             coeff, framework.Variable
diff --git a/python/paddle/fluid/contrib/mixed_precision/bf16/amp_lists.py b/python/paddle/fluid/contrib/mixed_precision/bf16/amp_lists.py
index 33694f4d127eed76fd84ba8b0074d1229f38d9a3..180e28ddab351499f7e3ecf65a8384a32bd29454 100644
--- a/python/paddle/fluid/contrib/mixed_precision/bf16/amp_lists.py
+++ b/python/paddle/fluid/contrib/mixed_precision/bf16/amp_lists.py
@@ -24,7 +24,7 @@ from ..fp16_lists import (
 __all__ = ["AutoMixedPrecisionListsBF16"]
 
 
-class AutoMixedPrecisionListsBF16(object):
+class AutoMixedPrecisionListsBF16:
     """
     AutoMixedPrecisionListsBF16 is a class for fp32/bf16 op types list. The lists are used for an
     algorithm which determines op's execution mode (fp32 or bf16).It can update pre-defined
diff --git a/python/paddle/fluid/contrib/mixed_precision/bf16/decorator.py b/python/paddle/fluid/contrib/mixed_precision/bf16/decorator.py
index 9110686582220fe1762e2c52551638f85d5ff828..dd1b07bfff6d5f1886c08260436a45d8a9056f77 100644
--- a/python/paddle/fluid/contrib/mixed_precision/bf16/decorator.py
+++ b/python/paddle/fluid/contrib/mixed_precision/bf16/decorator.py
@@ -31,7 +31,7 @@ import warnings
 __all__ = ["decorate_bf16"]
 
 
-class OptimizerWithMixedPrecision(object):
+class OptimizerWithMixedPrecision:
     """
     Optimizer with mixed-precision (MP) training. This is a wrapper of a common
     optimizer, plus the support of mixed-precision pre-training. The object
diff --git a/python/paddle/fluid/contrib/mixed_precision/decorator.py b/python/paddle/fluid/contrib/mixed_precision/decorator.py
index 75554ff9c812307364ae52c118005dd3724d356b..6b9f3f6eaabc51cf32845f6b704d2aa4ed7a9c14 100644
--- a/python/paddle/fluid/contrib/mixed_precision/decorator.py
+++ b/python/paddle/fluid/contrib/mixed_precision/decorator.py
@@ -34,7 +34,7 @@ import paddle
 __all__ = ["decorate"]
 
 
-class OptimizerWithMixedPrecision(object):
+class OptimizerWithMixedPrecision:
     """
     Optimizer with mixed-precision (MP) training. This is a wrapper of a common
     optimizer, plus the support of mixed-precision pre-training. The object
diff --git a/python/paddle/fluid/contrib/mixed_precision/fp16_lists.py b/python/paddle/fluid/contrib/mixed_precision/fp16_lists.py
index ef8f222bac23b22244748968385bbfb67898d01c..101af59861880898c4d5bdfb80536659f1648769 100644
--- a/python/paddle/fluid/contrib/mixed_precision/fp16_lists.py
+++ b/python/paddle/fluid/contrib/mixed_precision/fp16_lists.py
@@ -26,7 +26,7 @@ _extra_unsupported_fp16_list = {
 }
 
 
-class AutoMixedPrecisionLists(object):
+class AutoMixedPrecisionLists:
     """
     AutoMixedPrecisionLists is a class for black/white list. It can update
     pre-defined black list and white list according to users' custom black
diff --git a/python/paddle/fluid/contrib/quantize/quantize_transpiler.py b/python/paddle/fluid/contrib/quantize/quantize_transpiler.py
index 6e225fdbcc8e6c8731e730d831d6907adc1d7d77..edd07c0ba9f3684087ed7720c35e296c84f39b22 100644
--- a/python/paddle/fluid/contrib/quantize/quantize_transpiler.py
+++ b/python/paddle/fluid/contrib/quantize/quantize_transpiler.py
@@ -81,7 +81,7 @@ def quant(x, scale, num_bits):
     return y
 
 
-class QuantizeTranspiler(object):
+class QuantizeTranspiler:
     def __init__(
         self,
         weight_bits=8,
diff --git a/python/paddle/fluid/contrib/slim/quantization/adaround.py b/python/paddle/fluid/contrib/slim/quantization/adaround.py
index 9dd00ddadc6acd0eaf9af25d713ef564250ea960..278994ef318a12ef83c063147a88977b27ee2b96 100644
--- a/python/paddle/fluid/contrib/slim/quantization/adaround.py
+++ b/python/paddle/fluid/contrib/slim/quantization/adaround.py
@@ -51,7 +51,7 @@ def compute_soft_rounding_np(alpha_v):
     )
 
 
-class AdaRoundLoss(object):
+class AdaRoundLoss:
     def __init__(self, reg_param=0.01, default_beta_range=(20, 2)):
         self.default_reg_param = reg_param
         self.default_beta_range = default_beta_range
@@ -111,7 +111,7 @@ class AdaRoundLoss(object):
         return beta
 
 
-class AdaRound(object):
+class AdaRound:
     def __init__(
         self,
         scale,
diff --git a/python/paddle/fluid/contrib/slim/quantization/imperative/ptq.py b/python/paddle/fluid/contrib/slim/quantization/imperative/ptq.py
index 79f0a2bf359d737c8dfca8c77c2d70a32a7b85c7..358e08cf588903887086c0506ef6bc6e36c477bc 100644
--- a/python/paddle/fluid/contrib/slim/quantization/imperative/ptq.py
+++ b/python/paddle/fluid/contrib/slim/quantization/imperative/ptq.py
@@ -36,7 +36,7 @@ _logger = get_logger(
 )
 
 
-class ImperativePTQ(object):
+class ImperativePTQ:
     """
     Static post training quantization.
     """
diff --git a/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_config.py b/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_config.py
index b02de6cee4676ef2793a3dd066ead020e64ac704..88eb998c0e071ac7dc6a797e561e3ce349b1b099 100644
--- a/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_config.py
+++ b/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_config.py
@@ -22,7 +22,7 @@ from .ptq_quantizer import *
 __all__ = ['PTQConfig', 'default_ptq_config']
 
 
-class PTQConfig(object):
+class PTQConfig:
     """
     The PTQ config shows how to quantize the inputs and outputs.
     """
diff --git a/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_registry.py b/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_registry.py
index cb56bfcb47acbcef02728d9e77b33aaeffa929ab..d8df91f78fb8baadd9f82080934e4d2b19fc59d2 100644
--- a/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_registry.py
+++ b/python/paddle/fluid/contrib/slim/quantization/imperative/ptq_registry.py
@@ -17,7 +17,7 @@ import paddle
 __all__ = ['PTQRegistry']
 
 
-class LayerInfo(object):
+class LayerInfo:
     """
     Store the argnames of the inputs and outputs.
     """
@@ -63,7 +63,7 @@ QUANT_LAYERS_INFO = [
 SIMULATED_LAYERS = [paddle.nn.Conv2D, paddle.nn.Linear]
 
 
-class PTQRegistry(object):
+class PTQRegistry:
     """
     Register the supported layers for PTQ and provide layers info.
     """
diff --git a/python/paddle/fluid/contrib/slim/quantization/imperative/qat.py b/python/paddle/fluid/contrib/slim/quantization/imperative/qat.py
index aa768ab9745fe5332cf2ea7c5b1132edc3933763..9e64ecd96b11bd40e96580cd9432e19669b58408 100644
--- a/python/paddle/fluid/contrib/slim/quantization/imperative/qat.py
+++ b/python/paddle/fluid/contrib/slim/quantization/imperative/qat.py
@@ -57,7 +57,7 @@ def lazy_import_fleet(layer_name_map, fake_quant_input_layers):
     return layer_name_map, fake_quant_input_layers
 
 
-class ImperativeQuantAware(object):
+class ImperativeQuantAware:
     """
     Applying quantization aware training (QAT) to the dgraph model.
     """
@@ -304,7 +304,7 @@ class ImperativeQuantAware(object):
         )
 
 
-class ImperativeQuantizeInputs(object):
+class ImperativeQuantizeInputs:
     """
     Based on the input params, add the quant_dequant computational
     logic both for activation inputs and weight inputs.
@@ -448,7 +448,7 @@ class ImperativeQuantizeInputs(object):
         return quant_layers.__dict__[quant_layer_name](layer, **self._kwargs)
 
 
-class ImperativeQuantizeOutputs(object):
+class ImperativeQuantizeOutputs:
     """
     Calculate the output scales for target layers.
     """
diff --git a/python/paddle/fluid/contrib/slim/quantization/post_training_quantization.py b/python/paddle/fluid/contrib/slim/quantization/post_training_quantization.py
index 3db16060e0e609577f085e1182db69f0b78ca311..68df2a8adcc961c2b34a6cbc92d803fa1bb211c2 100644
--- a/python/paddle/fluid/contrib/slim/quantization/post_training_quantization.py
+++ b/python/paddle/fluid/contrib/slim/quantization/post_training_quantization.py
@@ -112,7 +112,7 @@ def _apply_pass(
     return graph
 
 
-class PostTrainingQuantization(object):
+class PostTrainingQuantization:
     """
     Utilizing post training quantization methon to quantize the FP32 model,
     and it uses calibrate data to get the quantization information for all
@@ -1481,7 +1481,7 @@ class PostTrainingQuantizationProgram(PostTrainingQuantization):
         self._fetch_list = fetch_list
 
 
-class WeightQuantization(object):
+class WeightQuantization:
     _supported_quantizable_op_type = ['conv2d', 'depthwise_conv2d', 'mul']
     _supported_weight_quantize_type = ['channel_wise_abs_max', 'abs_max']
 
diff --git a/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py b/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py
index fcc2daff208ef285795ba695610adfcaf6c0e3ef..a617bac359b6f10b1d642a72283c722ce95b3bb8 100644
--- a/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py
+++ b/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py
@@ -22,7 +22,7 @@ __all__ = ['Quant2Int8MkldnnPass']
 OpRole = core.op_proto_and_checker_maker.OpRole
 
 
-class Quant2Int8MkldnnPass(object):
+class Quant2Int8MkldnnPass:
     """
     Transform a quant model IrGraph into MKL-DNN supported INT8 IrGraph.
     The pass consists of the following transformations:
diff --git a/python/paddle/fluid/contrib/slim/quantization/quant_int8_mkldnn_pass.py b/python/paddle/fluid/contrib/slim/quantization/quant_int8_mkldnn_pass.py
index 73c611db0120600400a0e4a22118a859fe141032..25278fc6913ee9e0e122cddc65bc37313dcc8ebb 100644
--- a/python/paddle/fluid/contrib/slim/quantization/quant_int8_mkldnn_pass.py
+++ b/python/paddle/fluid/contrib/slim/quantization/quant_int8_mkldnn_pass.py
@@ -21,7 +21,7 @@ from ....framework import _get_paddle_place
 __all__ = ['QuantInt8MkldnnPass']
 
 
-class QuantInt8MkldnnPass(object):
+class QuantInt8MkldnnPass:
     """
     Convert QuantizationFreezePass generated IrGraph to MKL-DNN supported INT8
     IrGraph. Following transformations did in this pass:
diff --git a/python/paddle/fluid/contrib/slim/quantization/quantization_pass.py b/python/paddle/fluid/contrib/slim/quantization/quantization_pass.py
index 020bdcec489d6d28342df5ec468dce8fefdf624c..f0caabd6f4ea19a0ca22e64bd80323281b54aaa5 100644
--- a/python/paddle/fluid/contrib/slim/quantization/quantization_pass.py
+++ b/python/paddle/fluid/contrib/slim/quantization/quantization_pass.py
@@ -107,7 +107,7 @@ def _check_grandchild_op_node(op_node, grandchild_op_name):
     return False
 
 
-class QuantizationTransformPass(object):
+class QuantizationTransformPass:
     """
     Quantize the ops that have weights. Add quant and dequant ops for
     the quantized ops's inputs.
@@ -1068,7 +1068,7 @@ class QuantizationTransformPass(object):
         return is_skip
 
 
-class QuantizationFreezePass(object):
+class QuantizationFreezePass:
     def __init__(
         self,
         scope,
@@ -1444,7 +1444,7 @@ class QuantizationFreezePass(object):
         )
 
 
-class ConvertToInt8Pass(object):
+class ConvertToInt8Pass:
     def __init__(self, scope, place, quantizable_op_type=None):
         """
         Convert the weights into int8_t type.
@@ -1537,7 +1537,7 @@ class ConvertToInt8Pass(object):
         graph.safe_remove_nodes(all_unused_vars)
 
 
-class TransformForMobilePass(object):
+class TransformForMobilePass:
     def __init__(self):
         """
         This pass is used to convert the frozen graph for paddle-mobile execution.
@@ -1579,7 +1579,7 @@ class TransformForMobilePass(object):
         return graph
 
 
-class OutScaleForTrainingPass(object):
+class OutScaleForTrainingPass:
     def __init__(
         self,
         scope=None,
@@ -1745,7 +1745,7 @@ class OutScaleForTrainingPass(object):
         return "%s@scale" % (var_name)
 
 
-class OutScaleForInferencePass(object):
+class OutScaleForInferencePass:
     def __init__(self, scope=None):
         """
         This pass is used for setting output scales of some operators.
@@ -1815,7 +1815,7 @@ class OutScaleForInferencePass(object):
         return "%s@scale" % (var_name)
 
 
-class AddQuantDequantPass(object):
+class AddQuantDequantPass:
     """
     Quantize the ops that do not have weights, and add quant_dequant op for the
     quantized ops's inputs.
@@ -2087,7 +2087,7 @@ class AddQuantDequantPass(object):
         return quant_var_node, scale_out_node
 
 
-class InsertQuantizeLinear(object):
+class InsertQuantizeLinear:
     """
     Insert quantize_linear and dequantize_linear op before ops.
 
@@ -2664,7 +2664,7 @@ class QuantizationTransformPassV2(QuantizationTransformPass):
         return graph
 
 
-class AddQuantDequantPassV2(object):
+class AddQuantDequantPassV2:
     """
     Quantize the ops that do not have weights, and add quant_linear and dequant_linear
     op for the quantized ops's inputs. It is used in the new format of quantization.
@@ -2850,7 +2850,7 @@ class AddQuantDequantPassV2(object):
         return graph
 
 
-class ReplaceFakeQuantDequantPass(object):
+class ReplaceFakeQuantDequantPass:
     """
     replace quant-dequant ops with quantize_linear and dequantize_linear ops.
     """
@@ -2987,7 +2987,7 @@ class ReplaceFakeQuantDequantPass(object):
         return "%s@zero_point" % (var_name)
 
 
-class QuantWeightPass(object):
+class QuantWeightPass:
     """
     quant weights and remove weights input quantize_linear node. for example:
     `weight -> quant -> dequant -> conv2d` will be frozen into `weight -> dequant -> conv2d`,
@@ -3129,7 +3129,7 @@ class QuantWeightPass(object):
         tensor.set(array, self._place)
 
 
-class AddQuantDequantForInferencePass(object):
+class AddQuantDequantForInferencePass:
     """
     When export quant model, it will traverse to find the output of each op, and then insert the quant/dequant op after it.
     """
diff --git a/python/paddle/fluid/contrib/slim/quantization/quantize_transpiler_v2.py b/python/paddle/fluid/contrib/slim/quantization/quantize_transpiler_v2.py
index dbc6277a3bf10d0a54a13a237636fbab5530a61a..64bb1a6c45a1c846c2b6f669771712710c4af1a5 100644
--- a/python/paddle/fluid/contrib/slim/quantization/quantize_transpiler_v2.py
+++ b/python/paddle/fluid/contrib/slim/quantization/quantize_transpiler_v2.py
@@ -49,7 +49,7 @@ def load_variable_data(scope, var_name):
     return np.array(var_node.get_tensor())
 
 
-class QuantizeTranspilerV2(object):
+class QuantizeTranspilerV2:
     def __init__(
         self,
         weight_bits=8,
diff --git a/python/paddle/fluid/contrib/slim/quantization/utils.py b/python/paddle/fluid/contrib/slim/quantization/utils.py
index 9862772c64a9881428e0ae1d0fc2e1d7eb099522..5f5fc99b44c7dbd46fc7a5f2c8236c14dbb3d9a6 100644
--- a/python/paddle/fluid/contrib/slim/quantization/utils.py
+++ b/python/paddle/fluid/contrib/slim/quantization/utils.py
@@ -472,7 +472,7 @@ def l2_loss(gt, pred):
     return ((gt - pred) ** 2).mean()
 
 
-class tqdm(object):
+class tqdm:
     def __init__(self, total, bar_format='Loading|{bar}', ncols=80):
         self.total = total
         self.bar_format = bar_format
diff --git a/python/paddle/fluid/contrib/sparsity/asp.py b/python/paddle/fluid/contrib/sparsity/asp.py
index 11fc3ca259f5fc25ff86a00406857428ca2caeec..d2165def6f0cbf6d8268b49386933df8074c82c1 100644
--- a/python/paddle/fluid/contrib/sparsity/asp.py
+++ b/python/paddle/fluid/contrib/sparsity/asp.py
@@ -478,7 +478,7 @@ def prune_model(model, n=2, m=4, mask_algo='mask_1d', with_mask=True):
     )
 
 
-class ProgramASPInfo(object):
+class ProgramASPInfo:
     r"""
     ProgramASPInfo is a container to keep ASP relevant information of Pragrom. It contains three inner-variables:
     1. __mask_vars (Dictionary): Key is parameter's name and vaule is its corresponding sparse mask Variable object, which is created by `ASPHelper.create_mask_variables`.
@@ -516,7 +516,7 @@ class ProgramASPInfo(object):
         return self.__excluded_layers
 
 
-class ASPHelper(object):
+class ASPHelper:
     r"""
     ASPHelper is a collection of Auto SParsity (ASP) functions to enable
 
@@ -917,7 +917,7 @@ class ASPHelper(object):
                 )
 
 
-class OptimizerWithSparsityGuarantee(object):
+class OptimizerWithSparsityGuarantee:
     r"""
     OptimizerWithSparsityGuarantee is a wrapper to decorate `minimize` function of given optimizer by `_minimize` of ASPHelper.
     The decorated `minimize` function would do three things (exactly same as `ASPHelper._minimize`):
diff --git a/python/paddle/fluid/data_feed_desc.py b/python/paddle/fluid/data_feed_desc.py
index 7a58c5ef5e1b41da615d77d3f2851d5c501cdb2a..9e6257d96ef6afbbf8226f49a483bb0e9429d406 100644
--- a/python/paddle/fluid/data_feed_desc.py
+++ b/python/paddle/fluid/data_feed_desc.py
@@ -18,7 +18,7 @@ from google.protobuf import text_format
 __all__ = ['DataFeedDesc']
 
 
-class DataFeedDesc(object):
+class DataFeedDesc:
     """
     :api_attr: Static Graph
 
diff --git a/python/paddle/fluid/data_feeder.py b/python/paddle/fluid/data_feeder.py
index e84d41e56204c22e50625c83e79177d6f4ecb1e8..1f900eff0df87f7636bc5667b547ad61fa9525e7 100644
--- a/python/paddle/fluid/data_feeder.py
+++ b/python/paddle/fluid/data_feeder.py
@@ -221,7 +221,7 @@ def check_shape(
         check_dtype(shape.dtype, 'shape', expected_tensor_dtype, op_name)
 
 
-class DataToLoDTensorConverter(object):
+class DataToLoDTensorConverter:
     def __init__(self, place, lod_level, shape, dtype):
         self.place = place
         self.lod_level = lod_level
@@ -280,7 +280,7 @@ class DataToLoDTensorConverter(object):
         return t
 
 
-class BatchedTensorProvider(object):
+class BatchedTensorProvider:
     def __init__(self, feed_list, place, batch_size, generator, drop_last):
         self.place = place
         self.batch_size = batch_size
@@ -319,7 +319,7 @@ class BatchedTensorProvider(object):
             [c._reset() for c in self.converters]
 
 
-class DataFeeder(object):
+class DataFeeder:
     """
     :api_attr: Static Graph
 
diff --git a/python/paddle/fluid/dataloader/batch_sampler.py b/python/paddle/fluid/dataloader/batch_sampler.py
index 624754ae286ad3d84473e8f64bd8df9160169d56..5ac1c79d0cdf70c68e045b96826fc36dc8cd7be3 100644
--- a/python/paddle/fluid/dataloader/batch_sampler.py
+++ b/python/paddle/fluid/dataloader/batch_sampler.py
@@ -159,7 +159,7 @@ class BatchSampler(Sampler):
         return num_samples // self.batch_size
 
 
-class _InfiniteIterableSampler(object):
+class _InfiniteIterableSampler:
     def __init__(self, dataset, batch_size=1):
         assert isinstance(
             dataset, IterableDataset
diff --git a/python/paddle/fluid/dataloader/dataloader_iter.py b/python/paddle/fluid/dataloader/dataloader_iter.py
index 06464d528df9e60bf71fc861c0ab1e9f54d9c9db..c89c3e079f62401567fd37a12816af8803273b94 100644
--- a/python/paddle/fluid/dataloader/dataloader_iter.py
+++ b/python/paddle/fluid/dataloader/dataloader_iter.py
@@ -90,7 +90,7 @@ def _clear_loader():
 CleanupFuncRegistrar.register(_clear_loader)
 
 
-class _DataLoaderIterBase(object):
+class _DataLoaderIterBase:
     """
     Iterator implement of DataLoader, will load and feed mini-batch
     data by setting in given dataloader.
diff --git a/python/paddle/fluid/dataloader/dataset.py b/python/paddle/fluid/dataloader/dataset.py
index 5fe52196cd88f2c3a4c5df10f6487973e38bc991..04e03ec844aacf37e4f6e02ee9c2590cb97e7d56 100755
--- a/python/paddle/fluid/dataloader/dataset.py
+++ b/python/paddle/fluid/dataloader/dataset.py
@@ -26,7 +26,7 @@ __all__ = [
 ]
 
 
-class Dataset(object):
+class Dataset:
     """
     An abstract class to encapsulate methods and behaviors of datasets.
 
diff --git a/python/paddle/fluid/dataloader/fetcher.py b/python/paddle/fluid/dataloader/fetcher.py
index 8d5a908729a6824fc963457f1be85823d8d6b78a..b097a315c0c7353631fac76e3bdaca52d6ea2435 100644
--- a/python/paddle/fluid/dataloader/fetcher.py
+++ b/python/paddle/fluid/dataloader/fetcher.py
@@ -19,7 +19,7 @@ from collections.abc import Sequence, Mapping
 _WARNING_TO_LOG = True
 
 
-class _DatasetFetcher(object):
+class _DatasetFetcher:
     def __init__(self, dataset, auto_collate_batch, collate_fn, drop_last):
         self.dataset = dataset
         self.auto_collate_batch = auto_collate_batch
diff --git a/python/paddle/fluid/dataloader/sampler.py b/python/paddle/fluid/dataloader/sampler.py
index 3626ed63e521c12f8bdfd0ef3001823c4a0acac6..afd8fa7da025705ff83fe6d3da20ffa483406f3a 100644
--- a/python/paddle/fluid/dataloader/sampler.py
+++ b/python/paddle/fluid/dataloader/sampler.py
@@ -23,7 +23,7 @@ __all__ = [
 ]
 
 
-class Sampler(object):
+class Sampler:
     """
     An abstract class to encapsulate methods and behaviors of samplers.
 
diff --git a/python/paddle/fluid/dataloader/worker.py b/python/paddle/fluid/dataloader/worker.py
index 6016c04054d6fc0bbc1ce80bb77881959044c8a1..155208791eacd2f828dbaf1d314617bf7accd8cf 100644
--- a/python/paddle/fluid/dataloader/worker.py
+++ b/python/paddle/fluid/dataloader/worker.py
@@ -34,16 +34,16 @@ import queue
 __all__ = ['get_worker_info']
 
 
-class _IterableDatasetStopIteration(object):
+class _IterableDatasetStopIteration:
     def __init__(self, worker_id):
         self.worker_id = worker_id
 
 
-class _ResumeIteration(object):
+class _ResumeIteration:
     pass
 
 
-class _DatasetKind(object):
+class _DatasetKind:
     MAP = 0
     ITER = 1
 
@@ -63,7 +63,7 @@ class _DatasetKind(object):
             raise NotImplementedError("unknown Dataset kind {}".format(kind))
 
 
-class ParentWatchDog(object):
+class ParentWatchDog:
     def __init__(self):
         self._parent_pid = os.getppid()
         self._parent_alive = True
@@ -145,7 +145,7 @@ def get_worker_info():
     return _worker_info
 
 
-class WorkerInfo(object):
+class WorkerInfo:
     __initialized = False
 
     def __init__(self, **kwargs):
@@ -163,7 +163,7 @@ class WorkerInfo(object):
         return super().__setattr__(key, val)
 
 
-class _WorkerException(object):
+class _WorkerException:
     def __init__(self, worker_id, exc_info=None):
         self.worker_id = worker_id
         exc_info = exc_info or sys.exc_info()
diff --git a/python/paddle/fluid/dataset.py b/python/paddle/fluid/dataset.py
index 330ab0635d949ab37c31912a1c77e8642ecf7d28..b21550bcc3ad328e5497100bf5f89ced11c63e42 100644
--- a/python/paddle/fluid/dataset.py
+++ b/python/paddle/fluid/dataset.py
@@ -21,7 +21,7 @@ from ..utils import deprecated
 __all__ = ['DatasetFactory', 'InMemoryDataset', 'QueueDataset']
 
 
-class DatasetFactory(object):
+class DatasetFactory:
     """
     DatasetFactory is a factory which create dataset by its name,
     you can create "QueueDataset" or "InMemoryDataset", or "FileInstantDataset",
@@ -64,7 +64,7 @@ class DatasetFactory(object):
             )
 
 
-class DatasetBase(object):
+class DatasetBase:
     """Base dataset class."""
 
     def __init__(self):
diff --git a/python/paddle/fluid/device_worker.py b/python/paddle/fluid/device_worker.py
index 3cf41b3597fd10a04b82de1f215a968b773b631a..6b3359ac6131157bb24eab69c565587b0dc156f4 100644
--- a/python/paddle/fluid/device_worker.py
+++ b/python/paddle/fluid/device_worker.py
@@ -23,7 +23,7 @@ __all__ = [
 ]
 
 
-class DeviceWorker(object):
+class DeviceWorker:
     """
     DeviceWorker is an abstract class, which generates worker desc.
     This class is an inner class that we do computation logics within
@@ -695,7 +695,7 @@ class HeterSection(DeviceWorker):
         )
 
 
-class DeviceWorkerFactory(object):
+class DeviceWorkerFactory:
     def _create_device_worker(self, worker_type):
         classname = worker_type.capitalize()
         return globals()[classname]()
diff --git a/python/paddle/fluid/distributed/downpour.py b/python/paddle/fluid/distributed/downpour.py
index c710b7337ccb0a51c0d29c5da85ff851a0f71812..1f7e146c3f35b6abf4bb57255096fc81a7fd01b9 100644
--- a/python/paddle/fluid/distributed/downpour.py
+++ b/python/paddle/fluid/distributed/downpour.py
@@ -25,7 +25,7 @@ from paddle.fluid.distribute_lookup_table import (
 from google.protobuf import text_format
 
 
-class DownpourSGD(object):
+class DownpourSGD:
     r"""
     Distributed optimizer of downpour stochastic gradient descent
     Standard implementation of Google's Downpour SGD
diff --git a/python/paddle/fluid/distributed/fleet.py b/python/paddle/fluid/distributed/fleet.py
index 08b46ce04733ab3c82e3d2bc33ec99656fb95c0f..552dc91598ba470699a9513d879f44b760bd2f8e 100644
--- a/python/paddle/fluid/distributed/fleet.py
+++ b/python/paddle/fluid/distributed/fleet.py
@@ -18,7 +18,7 @@ from google.protobuf import text_format
 __all__ = ['Fleet']
 
 
-class Fleet(object):
+class Fleet:
     """ """
 
     def __init__(self):
diff --git a/python/paddle/fluid/distributed/helper.py b/python/paddle/fluid/distributed/helper.py
index 4a1643733393a3e808625c887c5cae7c82e41a6f..93cec48408367b436ca8398c4d5573f8ca3af7ca 100644
--- a/python/paddle/fluid/distributed/helper.py
+++ b/python/paddle/fluid/distributed/helper.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class FileSystem(object):
+class FileSystem:
     """
     A file system that support hadoop client desc.
 
@@ -54,7 +54,7 @@ class FileSystem(object):
         return self.fs_client
 
 
-class MPIHelper(object):
+class MPIHelper:
     """
     MPIHelper is a wrapper of mpi4py, support get_rank get_size etc.
     Args:
diff --git a/python/paddle/fluid/distributed/node.py b/python/paddle/fluid/distributed/node.py
index 793787d0fd7b681bbe88956cbffc07fdfb22f6bf..19fa5ec20b96dc229e43a8dedef27087fb797512 100644
--- a/python/paddle/fluid/distributed/node.py
+++ b/python/paddle/fluid/distributed/node.py
@@ -17,7 +17,7 @@ import ps_pb2 as pslib
 from functools import reduce
 
 
-class Server(object):
+class Server:
     """
     A Server basic class.
     """
@@ -26,7 +26,7 @@ class Server(object):
         pass
 
 
-class Worker(object):
+class Worker:
     """
     A Worker basic class.
     """
diff --git a/python/paddle/fluid/distributed/ps_instance.py b/python/paddle/fluid/distributed/ps_instance.py
index 370e1b19647245f56b4b1177781eebace060f263..91388b97b01aa8472eb5460c8b14ad55e0d60437 100644
--- a/python/paddle/fluid/distributed/ps_instance.py
+++ b/python/paddle/fluid/distributed/ps_instance.py
@@ -14,7 +14,7 @@
 from .helper import MPIHelper
 
 
-class PaddlePSInstance(object):
+class PaddlePSInstance:
     """
     PaddlePSInstance class is used to generate A instance of server or worker
     Args:
diff --git a/python/paddle/fluid/dygraph/amp/auto_cast.py b/python/paddle/fluid/dygraph/amp/auto_cast.py
index bb1acc7c09bfcc7610bb622c14cf0fb21dfa875f..721d10d76b128639fa4d1b124bc002f8df5138af 100644
--- a/python/paddle/fluid/dygraph/amp/auto_cast.py
+++ b/python/paddle/fluid/dygraph/amp/auto_cast.py
@@ -463,7 +463,7 @@ def amp_guard(
             tracer._amp_dtype = original_amp_dtype
 
 
-class StateDictHook(object):
+class StateDictHook:
     def __init__(self, save_dtype):
         self._save_dtype = save_dtype
 
diff --git a/python/paddle/fluid/dygraph/amp/loss_scaler.py b/python/paddle/fluid/dygraph/amp/loss_scaler.py
index c59588e9d03b611fe3240f6edca52b42df9d4e3f..0985237f516dac03d843cee687fbf8dfceb536c1 100644
--- a/python/paddle/fluid/dygraph/amp/loss_scaler.py
+++ b/python/paddle/fluid/dygraph/amp/loss_scaler.py
@@ -41,7 +41,7 @@ def _refresh_optimizer_state():
     return {"state": OptimizerState.INIT}
 
 
-class AmpScaler(object):
+class AmpScaler:
     """
     :api_attr: imperative
 
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/base_transformer.py b/python/paddle/fluid/dygraph/dygraph_to_static/base_transformer.py
index 044f9055e7a391227815229fc2adae873e31aab7..5f082acca5d401e02ecdfb6a722b1ee23693a80b 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/base_transformer.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/base_transformer.py
@@ -200,7 +200,7 @@ class ForLoopTuplePreTransformer(BaseTransformer):
         return [assign_node]
 
 
-class ForNodeVisitor(object):
+class ForNodeVisitor:
     """
     This class parses python for statement, get transformed 3 statement components of for node
     three key statements:
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/convert_call_func.py b/python/paddle/fluid/dygraph/dygraph_to_static/convert_call_func.py
index 2ac1d6403c0f9b220c1196de18e6704dde1e9b0b..72b6cc98839d69c83e04af3c11852b826b84ae6f 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/convert_call_func.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/convert_call_func.py
@@ -60,7 +60,7 @@ translator_logger = TranslatorLogger()
 CONVERSION_OPTIONS = "An attribute for a function that indicates conversion flags of the function in dynamic-to-static."
 
 
-class ConversionOptions(object):
+class ConversionOptions:
     """
     A container for conversion flags of a function in dynamic-to-static.
 
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/error.py b/python/paddle/fluid/dygraph/dygraph_to_static/error.py
index 7e05b5db891f3dd118398e05b304f21a74afdf80..f4a66100807fda87f46084c6d2b63e7f9ae41906 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/error.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/error.py
@@ -153,7 +153,7 @@ class TraceBackFrameRange(OriginInfo):
         return msg + '\n'.join(self.source_code)
 
 
-class SuggestionDict(object):
+class SuggestionDict:
     def __init__(self):
         # {(keywords): (suggestions)}
         self.suggestion_dict = {
@@ -174,7 +174,7 @@ class Dy2StKeyError(Exception):
     pass
 
 
-class ErrorData(object):
+class ErrorData:
     """
     Error data attached to an exception which is raised in un-transformed code.
     """
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/function_spec.py b/python/paddle/fluid/dygraph/dygraph_to_static/function_spec.py
index 039df7e84e433a3ac4c60c7558ebbac4d82f055f..cc77e05dad1174f43ee59af9f8d010d117d893be 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/function_spec.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/function_spec.py
@@ -30,7 +30,7 @@ from paddle.fluid.dygraph.dygraph_to_static.utils import func_to_source_code
 from paddle.fluid.dygraph.io import TranslatedLayer
 
 
-class FunctionSpec(object):
+class FunctionSpec:
     """
     Wrapper class for a function for class method.
     """
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/logging_utils.py b/python/paddle/fluid/dygraph/dygraph_to_static/logging_utils.py
index 6f73a23316d272aaf34ad3116133d139d1dae22b..7b004964a472ddf65e13b3fb3e06044202c2a9a5 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/logging_utils.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/logging_utils.py
@@ -36,7 +36,7 @@ def synchronized(func):
     return wrapper
 
 
-class TranslatorLogger(object):
+class TranslatorLogger:
     """
     class for Logging and debugging during the tranformation from dygraph to static graph.
     The object of this class is a singleton.
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/origin_info.py b/python/paddle/fluid/dygraph/dygraph_to_static/origin_info.py
index 4bfc73a32565fa5a3bfc2fd04ee58c70a850bc24..d6ff463a70d7cd546d16c3defcfa1be82f2ce18f 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/origin_info.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/origin_info.py
@@ -27,7 +27,7 @@ except:
     from collections import Sequence
 
 
-class Location(object):
+class Location:
     """
     Location information of source code.
     """
@@ -53,7 +53,7 @@ class Location(object):
         return (self.filepath, self.lineno)
 
 
-class OriginInfo(object):
+class OriginInfo:
     """
     Original information of source code.
     """
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/partial_program.py b/python/paddle/fluid/dygraph/dygraph_to_static/partial_program.py
index 6eb38302a5841454449c19a7a083772341e29fc0..bc371cc99c9174df08278cd66d7069af486a1ce9 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/partial_program.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/partial_program.py
@@ -45,7 +45,7 @@ from paddle.fluid.dygraph.amp.auto_cast import (
 from paddle import _legacy_C_ops
 
 
-class NestSequence(object):
+class NestSequence:
     """
     A wrapper class that easily to flatten and restore the nest structure of
     given sequence.
@@ -108,7 +108,7 @@ class NestSequence(object):
         return self.__input_list[item]
 
 
-class LazyInitialized(object):
+class LazyInitialized:
     """
     Descriptor to implement lazy initialization of property.
     """
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/program_translator.py b/python/paddle/fluid/dygraph/dygraph_to_static/program_translator.py
index be7dc08403a8341d59a4c9bfc852bb8d1b9e7416..443242563132760b0135e8c8da09edffe31f71d0 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/program_translator.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/program_translator.py
@@ -68,7 +68,7 @@ __all__ = ['ProgramTranslator', 'convert_to_static']
 MAX_TRACED_PROGRAM_COUNT = 10
 
 
-class FunctionCache(object):
+class FunctionCache:
     """
     Caches the transformed functions to avoid redundant conversions of the same function.
     """
@@ -158,7 +158,7 @@ def convert_to_static(function):
         return static_func
 
 
-class CacheKey(object):
+class CacheKey:
     """
     Cached key for ProgramCache.
     """
@@ -280,7 +280,7 @@ def unwrap_decorators(func):
     return decorators, cur
 
 
-class StaticFunction(object):
+class StaticFunction:
     """
     Wrapper class to Manage program conversion of decorated function.
 
@@ -847,7 +847,7 @@ def _verify_init_in_dynamic_mode(class_instance):
             )
 
 
-class HookHelper(object):
+class HookHelper:
     """
     Only For converting pre/post hooks operation in outermost layer while jit.save.
     Because hooks in sublayer have been processed automatically.
@@ -901,7 +901,7 @@ class HookHelper(object):
         return outputs
 
 
-class ConcreteProgram(object):
+class ConcreteProgram:
 
     __slots__ = [
         'inputs',
@@ -1037,7 +1037,7 @@ def _extract_indeed_params_buffers(class_instance):
     return params + buffers
 
 
-class ProgramCache(object):
+class ProgramCache:
     """
     Wrapper class for the program functions defined by dygraph function.
     """
@@ -1119,7 +1119,7 @@ def synchronized(func):
     return lock_func
 
 
-class ProgramTranslator(object):
+class ProgramTranslator:
     """
     Class to translate dygraph function into static graph function. The object
     of this class is a singleton.
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/static_analysis.py b/python/paddle/fluid/dygraph/dygraph_to_static/static_analysis.py
index bf07523e12dba245fd0a821d19e02c6f17d1a30f..5b6c3d1261c840100667a109f235f386529d1811 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/static_analysis.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/static_analysis.py
@@ -25,7 +25,7 @@ from .utils import (
 __all__ = ['AstNodeWrapper', 'NodeVarType', 'StaticAnalysisVisitor']
 
 
-class NodeVarType(object):
+class NodeVarType:
     """
     Enum class of python variable types. We have to know some variable types
     during compile time to transfer AST. For example, a string variable and a
@@ -112,7 +112,7 @@ class NodeVarType(object):
         return NodeVarType.UNKNOWN
 
 
-class AstNodeWrapper(object):
+class AstNodeWrapper:
     """
     Wrapper for python gast.node. We need a node wrapper because gast.node
     doesn't store all required information when we are transforming AST.
@@ -127,7 +127,7 @@ class AstNodeWrapper(object):
         self.node_var_type = {NodeVarType.UNKNOWN}
 
 
-class AstVarScope(object):
+class AstVarScope:
     """
     AstVarScope is a class holding the map from current scope variable to its
     type.
@@ -181,7 +181,7 @@ class AstVarScope(object):
         return self.parent_scope.get_var_type(var_name)
 
 
-class AstVarEnv(object):
+class AstVarEnv:
     """
     A class maintains scopes and mapping from name strings to type.
     """
@@ -231,7 +231,7 @@ class AstVarEnv(object):
         return cur_scope_dict
 
 
-class StaticAnalysisVisitor(object):
+class StaticAnalysisVisitor:
     """
     A class that does static analysis
     """
diff --git a/python/paddle/fluid/dygraph/io.py b/python/paddle/fluid/dygraph/io.py
index 383fb1eff5d4509ff566099ed71b88c03ab38481..eca171cacd3300e1e8863d507f7bab1913d18b41 100644
--- a/python/paddle/fluid/dygraph/io.py
+++ b/python/paddle/fluid/dygraph/io.py
@@ -325,7 +325,7 @@ def _change_is_test_status(program_desc, is_test):
                 op._set_attr('is_test', is_test)
 
 
-class _ProgramHolder(object):
+class _ProgramHolder:
     """
     Holds the execution information of a Program.
 
diff --git a/python/paddle/fluid/dygraph/jit.py b/python/paddle/fluid/dygraph/jit.py
index 9936b4a23ac92e87402ebe721296570154773f38..c359c6d152f9efbe46c06b312a85e6e6eb3d4f21 100644
--- a/python/paddle/fluid/dygraph/jit.py
+++ b/python/paddle/fluid/dygraph/jit.py
@@ -334,7 +334,7 @@ def not_to_static(func=None):
     return func
 
 
-class _SaveLoadConfig(object):
+class _SaveLoadConfig:
     def __init__(self):
         self._output_spec = None
         self._model_filename = None
@@ -621,7 +621,7 @@ _save_pre_hooks_lock = threading.Lock()
 _save_pre_hooks = []
 
 
-class HookRemoveHelper(object):
+class HookRemoveHelper:
     """A HookRemoveHelper that can be used to remove hook."""
 
     def __init__(self, hook):
@@ -1496,7 +1496,7 @@ def _trace(
     return original_outputs, program, feed_names, fetch_names, parameters
 
 
-class TracedLayer(object):
+class TracedLayer:
     """
     :api_attr: imperative
 
diff --git a/python/paddle/fluid/dygraph/layers.py b/python/paddle/fluid/dygraph/layers.py
index 5187f9ae7229528f974a3d25758f3eee73480ffb..752694b614af70a9317fa9cdbdb66002264cb0c5 100644
--- a/python/paddle/fluid/dygraph/layers.py
+++ b/python/paddle/fluid/dygraph/layers.py
@@ -89,7 +89,7 @@ def _addindent(string, indent):
     return s1[0] + '\n' + '\n'.join(s2)
 
 
-class HookRemoveHelper(object):
+class HookRemoveHelper:
     """A HookRemoveHelper that can be used to remove hook."""
 
     next_hook_id = 0
@@ -105,7 +105,7 @@ class HookRemoveHelper(object):
             del hooks[self._hook_id]
 
 
-class Layer(object):
+class Layer:
     """
     Dynamic graph Layer based on OOD, includes the parameters of the layer, the structure of the forward graph and so on.
 
diff --git a/python/paddle/fluid/dygraph/learning_rate_scheduler.py b/python/paddle/fluid/dygraph/learning_rate_scheduler.py
index 3ca63f505cb9a8b6538ee6c18836fd096a1ef411..3afe92cbc62342aa9e34d96fdcb53e22afbc413b 100644
--- a/python/paddle/fluid/dygraph/learning_rate_scheduler.py
+++ b/python/paddle/fluid/dygraph/learning_rate_scheduler.py
@@ -35,7 +35,7 @@ __all__ = [
 ]
 
 
-class LearningRateDecay(object):
+class LearningRateDecay:
     """
     Base class of learning rate decay
 
diff --git a/python/paddle/fluid/dygraph/parallel.py b/python/paddle/fluid/dygraph/parallel.py
index a317cf676e4aeddbd454fa941f24d524f4d0ff77..85c95c6b2b3e68b60cd1cc01ecaba5fa42964616 100644
--- a/python/paddle/fluid/dygraph/parallel.py
+++ b/python/paddle/fluid/dygraph/parallel.py
@@ -82,7 +82,7 @@ def prepare_context(strategy=None):
     return strategy
 
 
-class ParallelEnv(object):
+class ParallelEnv:
     """
     .. note::
         This API is not recommended, if you need to get rank and world_size,
diff --git a/python/paddle/fluid/dygraph/static_runner.py b/python/paddle/fluid/dygraph/static_runner.py
index afa2091caf7b03fd5b95ff6ab1fb52f883fb5569..c0a61fb3c2419e3ad10cf1a738bb081d21ff7c13 100644
--- a/python/paddle/fluid/dygraph/static_runner.py
+++ b/python/paddle/fluid/dygraph/static_runner.py
@@ -18,7 +18,7 @@ from paddle.fluid.dygraph.io import TranslatedLayer
 
 # NOTE: This class will be deprecated later.
 # It is kept here because PaddleHub is already using this API.
-class StaticModelRunner(object):
+class StaticModelRunner:
     """
     A Dynamic graph Layer for loading inference program and related parameters,
     and then performing fine-tune training or inference.
diff --git a/python/paddle/fluid/dygraph/varbase_patch_methods.py b/python/paddle/fluid/dygraph/varbase_patch_methods.py
index 51fb34a24e657dc830c5c3c3b41e566489611fac..7c7aa964cf84d5b2e1ee62a52cc04c0d8e0feda5 100644
--- a/python/paddle/fluid/dygraph/varbase_patch_methods.py
+++ b/python/paddle/fluid/dygraph/varbase_patch_methods.py
@@ -44,7 +44,7 @@ from paddle import _C_ops, _legacy_C_ops
 _grad_scalar = None
 
 
-class TensorHookRemoveHelper(object):
+class TensorHookRemoveHelper:
     """
     A helper class that for removing Tensor gradient's hook.
     NOTE(wuweilong):the operation weakref.ref(tensor) will cause some unexpected errors in eager mode.
diff --git a/python/paddle/fluid/entry_attr.py b/python/paddle/fluid/entry_attr.py
index a86f81e6dd997be68399f2186587b225306c4d78..29ab37a3f3f4cf5c57e4dbb8b96e973efea7a382 100644
--- a/python/paddle/fluid/entry_attr.py
+++ b/python/paddle/fluid/entry_attr.py
@@ -15,7 +15,7 @@
 __all__ = ['ProbabilityEntry', 'CountFilterEntry']
 
 
-class EntryAttr(object):
+class EntryAttr:
     """
     Examples:
         .. code-block:: python
diff --git a/python/paddle/fluid/evaluator.py b/python/paddle/fluid/evaluator.py
index a9d0c3c0594aec052cb41b72f0f9f1ed0ef11d8e..007337b9d0404fa264f7627ced35b5a3980a4701 100644
--- a/python/paddle/fluid/evaluator.py
+++ b/python/paddle/fluid/evaluator.py
@@ -41,7 +41,7 @@ def _clone_var_(block, var):
     )
 
 
-class Evaluator(object):
+class Evaluator:
     """
     Warning: better to use the fluid.metrics.* things, more
     flexible support via pure Python and Operator, and decoupled
diff --git a/python/paddle/fluid/executor.py b/python/paddle/fluid/executor.py
index 2dbf2d5cedafcb440bf275c616614fe85883686e..41440cfdb0e1a9408a03902a1bce1d29b4670c48 100755
--- a/python/paddle/fluid/executor.py
+++ b/python/paddle/fluid/executor.py
@@ -620,7 +620,7 @@ def _as_lodtensor(data, place, dtype=None):
     return tensor
 
 
-class FetchHandler(object):
+class FetchHandler:
     def __init__(self, var_dict=None, period_secs=60):
         assert var_dict is not None
         self.var_dict = var_dict
@@ -648,7 +648,7 @@ handler = FetchHandlerExample(var_dict=var_dict)
         )
 
 
-class _StandaloneExecutor(object):
+class _StandaloneExecutor:
     def __init__(self, place, main_program, scope):
         self._place = core.Place()
         self._place.set_place(place)
@@ -736,8 +736,8 @@ class _StandaloneExecutor(object):
         return res
 
 
-class _ExecutorCache(object):
-    class _CachedData(object):
+class _ExecutorCache:
+    class _CachedData:
         def __init__(
             self,
             program,
@@ -908,7 +908,7 @@ class _ExecutorCache(object):
         return new_program, new_exe
 
 
-class Executor(object):
+class Executor:
     """
     :api_attr: Static Graph
 
diff --git a/python/paddle/fluid/framework.py b/python/paddle/fluid/framework.py
index e68286844ed6769464a6b740fec5920c55d76603..9f30a4e08a31fcf0b2167af577b20d9a99c5da7f 100644
--- a/python/paddle/fluid/framework.py
+++ b/python/paddle/fluid/framework.py
@@ -1077,7 +1077,7 @@ def mlu_places(device_ids=None):
     return [core.MLUPlace(dev_id) for dev_id in device_ids]
 
 
-class NameScope(object):
+class NameScope:
     def __init__(self, name="", parent=None):
         self._children = dict()
         self._name = name
@@ -2655,7 +2655,7 @@ def get_all_op_protos():
     return ret_values
 
 
-class OpProtoHolder(object):
+class OpProtoHolder:
     """
     A global variable to hold all OpProtos from C++ as a map
     """
@@ -2709,7 +2709,7 @@ class OpProtoHolder(object):
         }
 
 
-class Operator(object):
+class Operator:
     """
     In Fluid, all the operation are represented by Operator, and Operator
     is regarded as a build in an instruction of a Block. Users can use the
@@ -3550,7 +3550,7 @@ class Operator(object):
         self.desc.dist_attr = dist_attr
 
 
-class Block(object):
+class Block:
     """
     In Fluid, a Program is consistence of multi-Block, and Block stores
     VarDesc and OpDesc. In a specific Block, a VarDesc have a unique name.
@@ -4342,7 +4342,7 @@ def _apply_pass(
     return attrs
 
 
-class IrNode(object):
+class IrNode:
     """
     Python IrNode. Beneath it is a core.Node, which is used for Ir Pass.
     """
@@ -4783,7 +4783,7 @@ class IrOpNode(IrNode):
         return [IrVarNode(n) for n in self.node.outputs]
 
 
-class IrGraph(object):
+class IrGraph:
     """
     Python IrGraph. Beneath it is a core.Graph, which is used for
     creating a c++ Ir Pass Graph. An IrGraph is just a graph view of
@@ -5218,7 +5218,7 @@ class IrGraph(object):
             desc._set_attr(name, val)
 
 
-class Program(object):
+class Program:
     """
     Create Python Program.  It has at least one :ref:`api_guide_Block_en`, when the
     control flow op like conditional_block, while :ref:`api_paddle_fluid_layers_While` is included,
diff --git a/python/paddle/fluid/graphviz.py b/python/paddle/fluid/graphviz.py
index b7ef1c7531fc00fd47167303500e64adf997f395..0f9bc17edfbf3ef863070593a6e3eadb168fa510 100644
--- a/python/paddle/fluid/graphviz.py
+++ b/python/paddle/fluid/graphviz.py
@@ -25,7 +25,7 @@ def crepr(v):
     return str(v)
 
 
-class Rank(object):
+class Rank:
     def __init__(self, kind, name, priority):
         '''
         kind: str
@@ -49,7 +49,7 @@ class Rank(object):
         )
 
 
-class Graph(object):
+class Graph:
     rank_counter = 0
 
     def __init__(self, title, **attrs):
@@ -145,7 +145,7 @@ class Graph(object):
         return '\n'.join(reprs)
 
 
-class Node(object):
+class Node:
     counter = 1
 
     def __init__(self, label, prefix, description="", **attrs):
@@ -170,7 +170,7 @@ class Node(object):
         return reprs
 
 
-class Edge(object):
+class Edge:
     def __init__(self, source, target, **attrs):
         '''
         Link source to target.
@@ -199,7 +199,7 @@ class Edge(object):
         return repr
 
 
-class GraphPreviewGenerator(object):
+class GraphPreviewGenerator:
     '''
     Generate a graph image for ONNX proto.
     '''
diff --git a/python/paddle/fluid/incubate/checkpoint/auto_checkpoint.py b/python/paddle/fluid/incubate/checkpoint/auto_checkpoint.py
index 5fa0ed085b1870c1cc6c6d62d8eedb5b753aa7bd..33c95f03b1f5a981848166a0140b50bb79cb5e83 100644
--- a/python/paddle/fluid/incubate/checkpoint/auto_checkpoint.py
+++ b/python/paddle/fluid/incubate/checkpoint/auto_checkpoint.py
@@ -69,7 +69,7 @@ def _thread_checker():
     ), "auto checkpoint must run under main thread"
 
 
-class AutoCheckpointChecker(object):
+class AutoCheckpointChecker:
     def __init__(self):
         self._run_env = None
         self._platform = None
diff --git a/python/paddle/fluid/incubate/checkpoint/checkpoint_saver.py b/python/paddle/fluid/incubate/checkpoint/checkpoint_saver.py
index 21e305afc16bca0dba630113f17ff17fcd6577ee..79161fe0fa0203165b8611c59493765ac988e53e 100644
--- a/python/paddle/fluid/incubate/checkpoint/checkpoint_saver.py
+++ b/python/paddle/fluid/incubate/checkpoint/checkpoint_saver.py
@@ -15,7 +15,7 @@
 from ...compiler import CompiledProgram
 
 
-class SerializableBase(object):
+class SerializableBase:
     def serialize(self, path):
         raise NotImplementedError
 
@@ -54,7 +54,7 @@ class PaddleModel(SerializableBase):
         )
 
 
-class CheckpointSaver(object):
+class CheckpointSaver:
     def __init__(self, fs):
         self._fs = fs
         self._checkpoint_prefix = "__paddle_checkpoint__"
diff --git a/python/paddle/fluid/incubate/data_generator/__init__.py b/python/paddle/fluid/incubate/data_generator/__init__.py
index 4729f44f2b12c65d16da3b69dd205a65adda3fcb..3e66b75e28faf82607d3baa6477c9364d2e6c096 100644
--- a/python/paddle/fluid/incubate/data_generator/__init__.py
+++ b/python/paddle/fluid/incubate/data_generator/__init__.py
@@ -18,7 +18,7 @@ import sys
 __all__ = ['MultiSlotDataGenerator', 'MultiSlotStringDataGenerator']
 
 
-class DataGenerator(object):
+class DataGenerator:
     """
     DataGenerator is a general Base class for user to inherit
     A user who wants to define his/her own python processing logic
diff --git a/python/paddle/fluid/incubate/fleet/base/fleet_base.py b/python/paddle/fluid/incubate/fleet/base/fleet_base.py
index ea03e9305605a19d8ad792180ef297f66784bba0..61b7fe7de50625fabbf1848fec2e96072d8d417e 100644
--- a/python/paddle/fluid/incubate/fleet/base/fleet_base.py
+++ b/python/paddle/fluid/incubate/fleet/base/fleet_base.py
@@ -30,7 +30,7 @@ __all__ = ['Fleet', 'DistributedOptimizer']
 __all__ += mode.__all__
 
 
-class Fleet(object):
+class Fleet:
     """
     Fleet is the base class, transpiler and pslib are implementation of Fleet.
 
@@ -268,7 +268,7 @@ class Fleet(object):
         pass
 
 
-class DistributedOptimizer(object):
+class DistributedOptimizer:
     """
     DistributedOptimizer is a wrapper for paddle.fluid.optimizer
     A user should pass a paddle.fluid.optimizer to DistributedOptimizer
diff --git a/python/paddle/fluid/incubate/fleet/base/role_maker.py b/python/paddle/fluid/incubate/fleet/base/role_maker.py
index d0e27dc8bdc83c36c39c89c116319a72ae9aacec..3f639d7753b13ac2b0e8e1a3aca8f68120d49d88 100644
--- a/python/paddle/fluid/incubate/fleet/base/role_maker.py
+++ b/python/paddle/fluid/incubate/fleet/base/role_maker.py
@@ -35,7 +35,7 @@ class Role:
     XPU = 3
 
 
-class MockBarrier(object):
+class MockBarrier:
     """
     MockBarrier is a empty impletation for barrier
     mock as a real barrier for never-barrier in a specific scenario
@@ -70,7 +70,7 @@ class MockBarrier(object):
         return [obj]
 
 
-class RoleMakerBase(object):
+class RoleMakerBase:
     """
     RoleMakerBase is a base class for assigning a role to current process
     in distributed training.
diff --git a/python/paddle/fluid/incubate/fleet/collective/__init__.py b/python/paddle/fluid/incubate/fleet/collective/__init__.py
index 51900ccbd90dd023114de50325ce4585e5030457..949ef93a472a30a0c85a38e46edbbaa05befdc81 100644
--- a/python/paddle/fluid/incubate/fleet/collective/__init__.py
+++ b/python/paddle/fluid/incubate/fleet/collective/__init__.py
@@ -40,12 +40,12 @@ import re
 import shutil
 
 
-class LambConfig(object):
+class LambConfig:
     def __init__(self):
         pass
 
 
-class DistFCConfig(object):
+class DistFCConfig:
     def __init__(self):
         pass
 
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/distribute_transpiler/distributed_strategy.py b/python/paddle/fluid/incubate/fleet/parameter_server/distribute_transpiler/distributed_strategy.py
index d2d32d4529efba95c92b0dc31e7ba38c6cfb2bbf..0958dafd3c2f88ce37049cfbe61624688d79922f 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/distribute_transpiler/distributed_strategy.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/distribute_transpiler/distributed_strategy.py
@@ -31,7 +31,7 @@ from paddle.fluid.transpiler.distribute_transpiler import (
 from paddle.fluid.incubate.fleet.parameter_server.mode import DistributedMode
 
 
-class TrainerRuntimeConfig(object):
+class TrainerRuntimeConfig:
     def __init__(self):
         self.mode = None
         num_threads = os.getenv("CPU_NUM", "1")
@@ -160,7 +160,7 @@ class TrainerRuntimeConfig(object):
         return self.display(self.get_communicator_flags())
 
 
-class PSLibRuntimeConfig(object):
+class PSLibRuntimeConfig:
     def __init__(self):
         self.runtime_configs = {}
 
@@ -168,7 +168,7 @@ class PSLibRuntimeConfig(object):
         return self.runtime_configs
 
 
-class DistributedStrategy(object):
+class DistributedStrategy:
     def __init__(self):
         self._program_config = DistributeTranspilerConfig()
         self._trainer_runtime_config = TrainerRuntimeConfig()
@@ -456,7 +456,7 @@ class GeoStrategy(DistributedStrategy):
         self._build_strategy.async_mode = True
 
 
-class StrategyFactory(object):
+class StrategyFactory:
     def __init_(self):
         pass
 
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/ir/ps_dispatcher.py b/python/paddle/fluid/incubate/fleet/parameter_server/ir/ps_dispatcher.py
index 441246879dfcf23c83badda6c347092fe4d7a3ca..4155413cb59f9b4745522ef86fe56aa83e7ebb75 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/ir/ps_dispatcher.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/ir/ps_dispatcher.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class PSDispatcher(object):
+class PSDispatcher:
     """
     PSDispatcher is the base class for dispatching vars
     into different pserver instance.
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/ir/public.py b/python/paddle/fluid/incubate/fleet/parameter_server/ir/public.py
index 865ffd2e0fd0c4c1184b56b844c4afbb25292d6c..53460cf036ff217892a504f00831ce2b2686a56f 100755
--- a/python/paddle/fluid/incubate/fleet/parameter_server/ir/public.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/ir/public.py
@@ -135,7 +135,7 @@ def Singleton(cls):
 
 
 @Singleton
-class CompileTimeStrategy(object):
+class CompileTimeStrategy:
     def __init__(self, main_program, startup_program, strategy, role_maker):
         self.min_block_size = 81920
 
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/ir/ufind.py b/python/paddle/fluid/incubate/fleet/parameter_server/ir/ufind.py
index fef6f24570c17bfc28dc87d699891d84c292a59e..f6a05cbf86c6b938e83dd4d4f7c7476fdcb5c1e6 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/ir/ufind.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/ir/ufind.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class UnionFind(object):
+class UnionFind:
     """Union-find data structure.
 
     Union-find is a data structure that keeps track of a set of elements partitioned
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/ir/vars_metatools.py b/python/paddle/fluid/incubate/fleet/parameter_server/ir/vars_metatools.py
index 745c05d986a6e66dbf4dce6089dbbe44d05cc84c..e32b1253d4f33c16105653cf4f4188549b4c361f 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/ir/vars_metatools.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/ir/vars_metatools.py
@@ -52,7 +52,7 @@ def create_var_struct(var):
     )
 
 
-class VarStruct(object):
+class VarStruct:
     """
     record part properties of a Variable in python.
     """
@@ -80,7 +80,7 @@ class VarStruct(object):
         )
 
 
-class VarDistributed(object):
+class VarDistributed:
     """
     a class to record the var distributed on parameter servers.
     the class will record the relationship between origin var and slice var.
@@ -187,7 +187,7 @@ class VarDistributed(object):
         )
 
 
-class VarsDistributed(object):
+class VarsDistributed:
     """
     a gather about VarDistributed with many methods to find distributed vars.
     through the class, we can get overview about the distributed parameters on parameter servers.
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/pslib/__init__.py b/python/paddle/fluid/incubate/fleet/parameter_server/pslib/__init__.py
index c9942581e011305acfad9650044c3350dd504348..8e5f39415feeedbaa50bdff8590e9e30f1d5e22b 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/pslib/__init__.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/pslib/__init__.py
@@ -1027,7 +1027,7 @@ def _fleet_embedding_v2(
     )
 
 
-class fleet_embedding(object):
+class fleet_embedding:
     """
     fleet embedding class, it is used as a wrapper
     Example:
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/pslib/node.py b/python/paddle/fluid/incubate/fleet/parameter_server/pslib/node.py
index d1c9fae5ccbeb368de74a6dd788995b2af0e3a28..73fcd18bdb7b732d280bd98087d65a5b44716c48 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/pslib/node.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/pslib/node.py
@@ -18,7 +18,7 @@ from . import ps_pb2 as pslib
 from functools import reduce
 
 
-class Server(object):
+class Server:
     """
     A Server basic class
     it's a base class, does not have implementation
@@ -28,7 +28,7 @@ class Server(object):
         pass
 
 
-class Worker(object):
+class Worker:
     """
     A Worker basic class.
     it's a base class, does not have implementation
diff --git a/python/paddle/fluid/incubate/fleet/parameter_server/pslib/optimizer_factory.py b/python/paddle/fluid/incubate/fleet/parameter_server/pslib/optimizer_factory.py
index b55431ba0f8a6923179c19ee32cb87738044a6bd..db5b90374ca324cc91c4dd1e0b3763141d18a519 100644
--- a/python/paddle/fluid/incubate/fleet/parameter_server/pslib/optimizer_factory.py
+++ b/python/paddle/fluid/incubate/fleet/parameter_server/pslib/optimizer_factory.py
@@ -54,7 +54,7 @@ ch.setFormatter(formatter)
 logger.addHandler(ch)
 
 
-class DistributedOptimizerImplBase(object):
+class DistributedOptimizerImplBase:
     """
     DistributedOptimizerImplBase
     base class of optimizers
diff --git a/python/paddle/fluid/incubate/fleet/utils/fleet_util.py b/python/paddle/fluid/incubate/fleet/utils/fleet_util.py
index 8218d941729ee535a6bd9a05d7f78f6fddf525d5..8d5203f201d084960e2c5098b97fef0b631c01b2 100644
--- a/python/paddle/fluid/incubate/fleet/utils/fleet_util.py
+++ b/python/paddle/fluid/incubate/fleet/utils/fleet_util.py
@@ -39,7 +39,7 @@ _logger = get_logger(
 fleet = None
 
 
-class FleetUtil(object):
+class FleetUtil:
     """
     FleetUtil provides some common functions for users' convenience.
 
diff --git a/python/paddle/fluid/incubate/fleet/utils/http_server.py b/python/paddle/fluid/incubate/fleet/utils/http_server.py
index 10e15dad0d804384321b20553724005906989bdf..79b436ff40c8323de4bfe29324ab3b7de94a25ff 100644
--- a/python/paddle/fluid/incubate/fleet/utils/http_server.py
+++ b/python/paddle/fluid/incubate/fleet/utils/http_server.py
@@ -126,7 +126,7 @@ class KVHandler(SimpleHTTPServer.SimpleHTTPRequestHandler):
         self.end_headers()
 
 
-class KVHTTPServer(HTTPServer, object):
+class KVHTTPServer(HTTPServer):
     """
     it is a http server storing kv pairs.
     """
diff --git a/python/paddle/fluid/initializer.py b/python/paddle/fluid/initializer.py
index bb162573e3a80d6b1be6d8db3f791c6ed10ddd8a..b0c573936b8160fdf0446c6235f87666a8a6bb09 100644
--- a/python/paddle/fluid/initializer.py
+++ b/python/paddle/fluid/initializer.py
@@ -55,7 +55,7 @@ _global_weight_initializer_ = None
 _global_bias_initializer_ = None
 
 
-class Initializer(object):
+class Initializer:
     """Base class for variable initializers
 
     Defines the common interface of variable initializers.
diff --git a/python/paddle/fluid/io.py b/python/paddle/fluid/io.py
index 789c9d0e3187be14a91df262c99b49f15f4d21dd..a308d5e261393424b71b1979769e58ea80ffb2f2 100644
--- a/python/paddle/fluid/io.py
+++ b/python/paddle/fluid/io.py
@@ -87,7 +87,7 @@ _logger = get_logger(
 )
 
 
-class _open_buffer(object):
+class _open_buffer:
     def __init__(self, buffer):
         self.buffer = buffer
 
diff --git a/python/paddle/fluid/ir.py b/python/paddle/fluid/ir.py
index 7dbe815b5e0e01e987eea838302212818e1b1cb6..fb077ed8b5f0dd46a7c3f8d3a97fa97a621c751e 100644
--- a/python/paddle/fluid/ir.py
+++ b/python/paddle/fluid/ir.py
@@ -140,7 +140,7 @@ def apply_build_strategy(
     return build_strategy
 
 
-class RegisterPassHelper(object):
+class RegisterPassHelper:
     _register_helpers = list()
 
     def __init__(self, pass_pairs, pass_type=str(), input_specs=dict()):
@@ -286,8 +286,8 @@ class RegisterPassHelper(object):
         return multi_pass_desc.SerializeToString()
 
 
-class PassDesc(object):
-    class AttrHelper(object):
+class PassDesc:
+    class AttrHelper:
         def __init__(self, obj, name, element_index=None):
             self._obj = obj
             self._name = name
@@ -422,7 +422,7 @@ class PassDesc(object):
                 self._attrs[name] = attr
             return attr
 
-    class OpHelper(object):
+    class OpHelper:
         def __init__(self, type=None):
             self._type = type
 
diff --git a/python/paddle/fluid/layer_helper_base.py b/python/paddle/fluid/layer_helper_base.py
index 1ea8d504add89abd78ba385cd161474ac3bc7bb7..91ec751cc282cc0341e22150103f5710dc320658 100644
--- a/python/paddle/fluid/layer_helper_base.py
+++ b/python/paddle/fluid/layer_helper_base.py
@@ -31,7 +31,7 @@ from .initializer import _global_weight_initializer, _global_bias_initializer
 __all__ = ['LayerHelperBase']
 
 
-class LayerHelperBase(object):
+class LayerHelperBase:
     # global dtype
     __dtype = "float32"
 
diff --git a/python/paddle/fluid/layers/control_flow.py b/python/paddle/fluid/layers/control_flow.py
index 13934b76afd307a0c9aee669cd8628e7df81ea2b..91065258a13ad89808c607928fac6911b36946b6 100755
--- a/python/paddle/fluid/layers/control_flow.py
+++ b/python/paddle/fluid/layers/control_flow.py
@@ -536,7 +536,7 @@ def Assert(cond, data=None, summarize=20, name=None):
     return op
 
 
-class BlockGuard(object):
+class BlockGuard:
     """
     BlockGuard class.
 
@@ -584,7 +584,7 @@ class BlockGuardWithCompletion(BlockGuard):
         return super().__exit__(exc_type, exc_val, exc_tb)
 
 
-class StaticRNNMemoryLink(object):
+class StaticRNNMemoryLink:
     """
     StaticRNNMemoryLink class.
 
@@ -607,7 +607,7 @@ class StaticRNNMemoryLink(object):
         self.mem = mem
 
 
-class StaticRNN(object):
+class StaticRNN:
     """
     :api_attr: Static Graph
 
@@ -1179,7 +1179,7 @@ def get_inputs_outputs_in_block(
     return inner_inputs, inner_outputs
 
 
-class While(object):
+class While:
     """
     :api_attr: Static Graph
 
@@ -2463,7 +2463,7 @@ class ConditionalBlockGuard(BlockGuard):
         return super().__exit__(exc_type, exc_val, exc_tb)
 
 
-class ConditionalBlock(object):
+class ConditionalBlock:
     '''
     **ConditionalBlock**
 
@@ -3173,7 +3173,7 @@ def case(pred_fn_pairs, default=None, name=None):
     return final_fn()
 
 
-class Switch(object):
+class Switch:
     """
     :api_attr: Static Graph
 
@@ -3303,7 +3303,7 @@ class Switch(object):
         return True
 
 
-class IfElseBlockGuard(object):
+class IfElseBlockGuard:
     def __init__(self, is_true, ifelse):
         if not isinstance(ifelse, IfElse):
             raise TypeError("ifelse must be an instance of IfElse class")
@@ -3340,7 +3340,7 @@ class IfElseBlockGuard(object):
         self.ie.status = IfElse.OUT_IF_ELSE_BLOCKS
 
 
-class IfElse(object):
+class IfElse:
     """
     :api_attr: Static Graph
 
@@ -3530,7 +3530,7 @@ class IfElse(object):
         return rlist
 
 
-class DynamicRNN(object):
+class DynamicRNN:
     """
     :api_attr: Static Graph
 
diff --git a/python/paddle/fluid/layers/distributions.py b/python/paddle/fluid/layers/distributions.py
index 174b2f8e36c08dc65f8641f27dd403661a4318b5..bd0b39caf0d9e5094129319c224c39ad254138b0 100644
--- a/python/paddle/fluid/layers/distributions.py
+++ b/python/paddle/fluid/layers/distributions.py
@@ -31,7 +31,7 @@ from ..data_feeder import (
 __all__ = ['Uniform', 'Normal', 'Categorical', 'MultivariateNormalDiag']
 
 
-class Distribution(object):
+class Distribution:
     """
     Distribution is the abstract base class for probability distributions.
     """
diff --git a/python/paddle/fluid/layers/io.py b/python/paddle/fluid/layers/io.py
index e4f765dcc663056cd7b59907aaa116ebd2e5f4c1..1dcc07a20d06c0beab05d8d0d1316543e50f3269 100644
--- a/python/paddle/fluid/layers/io.py
+++ b/python/paddle/fluid/layers/io.py
@@ -172,7 +172,7 @@ class BlockGuardServ(BlockGuard):
         return super().__exit__(exc_type, exc_val, exc_tb)
 
 
-class ListenAndServ(object):
+class ListenAndServ:
     """
     **ListenAndServ Layer**
 
diff --git a/python/paddle/fluid/layers/nn.py b/python/paddle/fluid/layers/nn.py
index 5623413af01fedd994438094c44932931b67a644..57e8a24e0dc75311703d417ad6c247f870679629 100644
--- a/python/paddle/fluid/layers/nn.py
+++ b/python/paddle/fluid/layers/nn.py
@@ -14771,7 +14771,7 @@ def temporal_shift(x, seg_num, shift_ratio=0.25, name=None, data_format="NCHW"):
     )
 
 
-class PyFuncRegistry(object):
+class PyFuncRegistry:
     _register_funcs = []
 
     def __init__(self, func):
diff --git a/python/paddle/fluid/layers/rnn.py b/python/paddle/fluid/layers/rnn.py
index 797272510f0c128409e51c77489e8a7ca09c6f77..5cfc6e37ce1cc5a8f2b7b6140dda43efa190f1bf 100644
--- a/python/paddle/fluid/layers/rnn.py
+++ b/python/paddle/fluid/layers/rnn.py
@@ -62,7 +62,7 @@ __all__ = [
 ]
 
 
-class RNNCell(object):
+class RNNCell:
     """
         :api_attr: Static Graph
 
@@ -160,7 +160,7 @@ class RNNCell(object):
                 return True
             return isinstance(seq, Sequence) and not isinstance(seq, str)
 
-        class Shape(object):
+        class Shape:
             def __init__(self, shape):
                 self.shape = shape if shape[0] == -1 else ([-1] + list(shape))
 
@@ -544,7 +544,7 @@ def rnn(
         )
 
 
-class ArrayWrapper(object):
+class ArrayWrapper:
     def __init__(self, x):
         self.array = [x]
 
@@ -823,7 +823,7 @@ def birnn(
     return outputs, final_states
 
 
-class Decoder(object):
+class Decoder:
     """
         :api_attr: Static Graph
 
@@ -1869,7 +1869,7 @@ def dynamic_decode(
         )
 
 
-class DecodeHelper(object):
+class DecodeHelper:
     """
     DecodeHelper is the base class for any helper instance used in `BasicDecoder`.
     It provides interface to implement sampling and produce inputs for the next
diff --git a/python/paddle/fluid/lazy_init.py b/python/paddle/fluid/lazy_init.py
index d2118259d03cf902d687199ba272b0501533507a..6242ad2c4eded0de908913630a11d6351a7c65a0 100644
--- a/python/paddle/fluid/lazy_init.py
+++ b/python/paddle/fluid/lazy_init.py
@@ -17,7 +17,7 @@ from . import framework
 __all__ = ["LazyGuard"]
 
 
-class LazyInitHelper(object):
+class LazyInitHelper:
     """
     A Helper Context to trigger switching mode between dygraph and static mode,
     and holds the startup program resource.
@@ -88,7 +88,7 @@ def lazy_init_helper():
     return _lazy_init_helper
 
 
-class LazyGuard(object):
+class LazyGuard:
     """
     LazyGuard is a wrapper interface for nn.Layer, it forwards the construct
     process of user defined Layer. Meanwhile, it provides necessary API to
diff --git a/python/paddle/fluid/metrics.py b/python/paddle/fluid/metrics.py
index 1f5d941dc0f66ba2f061dd4b2cd3c32492a7059e..5776b4efdf3b905bcc8a5d5bc1ec7d819d14ceea 100644
--- a/python/paddle/fluid/metrics.py
+++ b/python/paddle/fluid/metrics.py
@@ -55,7 +55,7 @@ def _is_number_or_matrix_(var):
     return _is_number_(var) or isinstance(var, np.ndarray)
 
 
-class MetricBase(object):
+class MetricBase:
     """
     In many cases, we usually have to split the test data into mini-batches for evaluating
     deep neural networks, therefore we need to collect the evaluation results of each
@@ -818,7 +818,7 @@ class Auc(MetricBase):
         )
 
 
-class DetectionMAP(object):
+class DetectionMAP:
     """
     Calculate the detection mean average precision (mAP).
 
diff --git a/python/paddle/fluid/op.py b/python/paddle/fluid/op.py
index c8b118127cc13b6a3b017716151f68c4b36a82ed..515a5f8e770a14a04d3c165990e34c86c9fe3f1c 100644
--- a/python/paddle/fluid/op.py
+++ b/python/paddle/fluid/op.py
@@ -35,7 +35,7 @@ def is_str(s):
     return isinstance(s, str)
 
 
-class OpDescCreationMethod(object):
+class OpDescCreationMethod:
     """
     Convert the user's input(only keyword arguments are supported) to OpDesc
     based on the OpProto.
@@ -181,7 +181,7 @@ class OpDescCreationMethod(object):
         return False
 
 
-class OpInfo(object):
+class OpInfo:
     def __init__(self, name, method, inputs, outputs, attrs, extra_attrs):
         self.name = name
         self.method = method
@@ -213,7 +213,7 @@ def create_op_creation_method(op_proto):
     )
 
 
-class OperatorFactory(object):
+class OperatorFactory:
     def __init__(self):
         self.op_methods = dict()
 
@@ -266,7 +266,7 @@ class OperatorFactory(object):
         return self.get_op_info(type).extra_attrs
 
 
-class __RecurrentOp__(object):
+class __RecurrentOp__:
     __proto__ = None
     type = "recurrent"
 
@@ -287,7 +287,7 @@ class __RecurrentOp__(object):
         return core.RecurrentOp.create(proto.SerializeToString())
 
 
-class __DynamicRecurrentOp__(object):
+class __DynamicRecurrentOp__:
     __proto__ = None
     type = "dynamic_recurrent"
 
@@ -308,7 +308,7 @@ class __DynamicRecurrentOp__(object):
         return core.DynamicRecurrentOp.create(proto.SerializeToString())
 
 
-class __CondOp__(object):
+class __CondOp__:
     __proto__ = None
     type = "cond"
 
diff --git a/python/paddle/fluid/optimizer.py b/python/paddle/fluid/optimizer.py
index 9f361827f06036b83cefcda737cd82e28f98a199..7eaa38636c2529fb2c0098443bb24c6ebf7dd9b6 100755
--- a/python/paddle/fluid/optimizer.py
+++ b/python/paddle/fluid/optimizer.py
@@ -99,7 +99,7 @@ __all__ = [
 ]
 
 
-class Optimizer(object):
+class Optimizer:
     """Optimizer Base class.
 
     Define the common interface of an optimizer.
@@ -4617,7 +4617,7 @@ class ModelAverage(Optimizer):
         executor.run(self.restore_program)
 
 
-class ExponentialMovingAverage(object):
+class ExponentialMovingAverage:
     r"""
         :api_attr: Static Graph
 
@@ -4877,7 +4877,7 @@ class ExponentialMovingAverage(object):
         executor.run(self.restore_program)
 
 
-class PipelineOptimizer(object):
+class PipelineOptimizer:
     """
         :api_attr: Static Graph
 
@@ -7600,7 +7600,7 @@ class RecomputeOptimizer(Optimizer):
         return optimize_ops, params_grads
 
 
-class LookaheadOptimizer(object):
+class LookaheadOptimizer:
     r"""
         :api_attr: Static Graph
 
@@ -7780,7 +7780,7 @@ class LookaheadOptimizer(object):
         return mini_out
 
 
-class GradientMergeOptimizer(object):
+class GradientMergeOptimizer:
     """
     Gradient Merge, also called as Gradient Accumulation,
     is a training strategy for larger batches. With this strategy,
diff --git a/python/paddle/fluid/parallel_executor.py b/python/paddle/fluid/parallel_executor.py
index c0a45b041e6a5d65c1deb18567d928dd72902236..269172ea994e22ff6a6d34da7d97bfd22f113311 100644
--- a/python/paddle/fluid/parallel_executor.py
+++ b/python/paddle/fluid/parallel_executor.py
@@ -25,7 +25,7 @@ ExecutionStrategy = core.ParallelExecutor.ExecutionStrategy
 BuildStrategy = core.ParallelExecutor.BuildStrategy
 
 
-class ParallelExecutor(object):
+class ParallelExecutor:
     """
         :api_attr: Static Graph
 
diff --git a/python/paddle/fluid/param_attr.py b/python/paddle/fluid/param_attr.py
index e1c8568d93a1eee655a4981ee58850bc21f53bc8..f251a654a992b58fab2beec6bb6ee0ba1817db37 100644
--- a/python/paddle/fluid/param_attr.py
+++ b/python/paddle/fluid/param_attr.py
@@ -22,7 +22,7 @@ __all__ = [
 ]
 
 
-class ParamAttr(object):
+class ParamAttr:
     """
 
     Note:
diff --git a/python/paddle/fluid/reader.py b/python/paddle/fluid/reader.py
index 83aedd052a838e2de8c80a3d973d02333ba29d8e..2ad3e5903aab55e99c337bc234e2a869c87b9ca1 100644
--- a/python/paddle/fluid/reader.py
+++ b/python/paddle/fluid/reader.py
@@ -145,7 +145,7 @@ def _reader_process_loop(batch_reader, data_queue):
         raise
 
 
-class DataLoaderBase(object):
+class DataLoaderBase:
     def __init__(self):
         self._places = None
 
@@ -181,7 +181,7 @@ class DataLoaderBase(object):
         return arr
 
 
-class AuToTune(object):
+class AuToTune:
     def __init__(self, loader):
         self.loader = loader
         self.max_num_worker = multiprocessing.cpu_count() / 2
@@ -318,7 +318,7 @@ class AuToTune(object):
         return best_workers
 
 
-class DataLoader(object):
+class DataLoader:
     """
     DataLoader prodives an iterator which iterates given dataset
     once by the batch_sampler.
diff --git a/python/paddle/fluid/regularizer.py b/python/paddle/fluid/regularizer.py
index 60112d182e24f5b9338f1760979b365a6555aaf9..1152f0bbd86b6aafeda6ed4f0def651d702ba455 100644
--- a/python/paddle/fluid/regularizer.py
+++ b/python/paddle/fluid/regularizer.py
@@ -22,7 +22,7 @@ from paddle import _C_ops, _legacy_C_ops
 __all__ = ['L1Decay', 'L2Decay', 'L1DecayRegularizer', 'L2DecayRegularizer']
 
 
-class WeightDecayRegularizer(object):
+class WeightDecayRegularizer:
     """Base class for weight decay regularizers
 
     Defines the common interface of weight-decay regularizers.
diff --git a/python/paddle/fluid/tests/unittests/collective/fleet/hybrid_parallel_communicate_group.py b/python/paddle/fluid/tests/unittests/collective/fleet/hybrid_parallel_communicate_group.py
index 0c6bf6e9deccef5df5341d260c1cc7afbab3f886..20e4b8312d08df22af03a7d795170411d47c07e3 100644
--- a/python/paddle/fluid/tests/unittests/collective/fleet/hybrid_parallel_communicate_group.py
+++ b/python/paddle/fluid/tests/unittests/collective/fleet/hybrid_parallel_communicate_group.py
@@ -17,7 +17,7 @@ import paddle
 from paddle.distributed import fleet
 
 
-class TestNewGroupAPI(object):
+class TestNewGroupAPI:
     def __init__(self):
         paddle.distributed.init_parallel_env()
         topo = fleet.CommunicateTopology(
diff --git a/python/paddle/fluid/tests/unittests/collective/fleet/new_group.py b/python/paddle/fluid/tests/unittests/collective/fleet/new_group.py
index 9fa469ea5e234aff2e60b3d5019ac544a9e82aad..40924c66bb4279981442f98024beebcb5e700100 100644
--- a/python/paddle/fluid/tests/unittests/collective/fleet/new_group.py
+++ b/python/paddle/fluid/tests/unittests/collective/fleet/new_group.py
@@ -16,7 +16,7 @@ import numpy as np
 import paddle
 
 
-class TestNewGroupAPI(object):
+class TestNewGroupAPI:
     def __init__(self):
         paddle.distributed.init_parallel_env()
         d1 = np.array([1, 2, 3])
diff --git a/python/paddle/fluid/tests/unittests/collective/fleet/parallel_dygraph_transformer.py b/python/paddle/fluid/tests/unittests/collective/fleet/parallel_dygraph_transformer.py
index c4c10f75a610a0bbf29b8ecb61f7757eb0b83fb9..f5b5903831b24170fe9674c3e905d29db086b3f9 100644
--- a/python/paddle/fluid/tests/unittests/collective/fleet/parallel_dygraph_transformer.py
+++ b/python/paddle/fluid/tests/unittests/collective/fleet/parallel_dygraph_transformer.py
@@ -36,7 +36,7 @@ Note(chenweihang): To compare loss of single-card and multi-card
 """
 
 
-class TrainTaskConfig(object):
+class TrainTaskConfig:
     """
     TrainTaskConfig
     """
@@ -61,7 +61,7 @@ class TrainTaskConfig(object):
     label_smooth_eps = 0.1
 
 
-class ModelHyperParams(object):
+class ModelHyperParams:
     # These following five vocabularies related configurations will be set
     # automatically according to the passed vocabulary path and special tokens.
     # size of source word dictionary.
diff --git a/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision.py b/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision.py
index 798847ed3e4a11d609b3542d03433ec44424fd70..c2ac727156420a93809edf36a29d67acdd55b6eb 100644
--- a/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision.py
+++ b/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision.py
@@ -612,7 +612,7 @@ class TestAmpDecorator(unittest.TestCase):
 
     def test_input_type_exception(self):
         def test_error_model():
-            class MyModel(object):
+            class MyModel:
                 def __init__(self):
                     print("A fake Model")
 
@@ -631,7 +631,7 @@ class TestAmpDecorator(unittest.TestCase):
         self.assertRaises(RuntimeError, test_error_distributed_model)
 
         def test_error_optimizer():
-            class MyOptimizer(object):
+            class MyOptimizer:
                 def __init__(self):
                     print("A fake Optimizer")
 
diff --git a/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision_for_eager.py b/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision_for_eager.py
index 7957cceb7b8a4c58080940782505829d5707ee78..e2955483935afc71496c1350cd6a415a0a1dfe69 100644
--- a/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision_for_eager.py
+++ b/python/paddle/fluid/tests/unittests/collective/fleet/test_imperative_auto_mixed_precision_for_eager.py
@@ -611,7 +611,7 @@ class TestAmpDecorator(unittest.TestCase):
 
     def test_input_type_exception(self):
         def test_error_model():
-            class MyModel(object):
+            class MyModel:
                 def __init__(self):
                     print("A fake Model")
 
@@ -630,7 +630,7 @@ class TestAmpDecorator(unittest.TestCase):
         self.assertRaises(RuntimeError, test_error_distributed_model)
 
         def test_error_optimizer():
-            class MyOptimizer(object):
+            class MyOptimizer:
                 def __init__(self):
                     print("A fake Optimizer")
 
diff --git a/python/paddle/fluid/tests/unittests/collective/multinode/test_collective_multi_nodes.py b/python/paddle/fluid/tests/unittests/collective/multinode/test_collective_multi_nodes.py
index 868585f81a0bede0809ac4a13336be3146b17dc7..786b81a06e9b58c4169a421fb0d54169eb91e755 100644
--- a/python/paddle/fluid/tests/unittests/collective/multinode/test_collective_multi_nodes.py
+++ b/python/paddle/fluid/tests/unittests/collective/multinode/test_collective_multi_nodes.py
@@ -19,7 +19,7 @@ import subprocess
 import tempfile
 
 
-class TestCollectiveAPIRunnerBase(object):
+class TestCollectiveAPIRunnerBase:
     def check_pass(self, *args, **kwargs):
         raise NotImplementedError(
             "get model should be implemented by child class."
diff --git a/python/paddle/fluid/tests/unittests/ctr_dataset_reader.py b/python/paddle/fluid/tests/unittests/ctr_dataset_reader.py
index 5cc96cf1c4e58cc757b9baf61f14e012eb4ea2f0..88d1327eaf390e57af1763355af2ae37dd963c64 100644
--- a/python/paddle/fluid/tests/unittests/ctr_dataset_reader.py
+++ b/python/paddle/fluid/tests/unittests/ctr_dataset_reader.py
@@ -60,7 +60,7 @@ def load_lr_input_record(sent):
     return res
 
 
-class CtrReader(object):
+class CtrReader:
     def __init__(self):
         pass
 
diff --git a/python/paddle/fluid/tests/unittests/dist_ctr_reader.py b/python/paddle/fluid/tests/unittests/dist_ctr_reader.py
index fafb5f19c5f7d062759a84e3cebed4b5e5a4bcbd..2c6fa74b363f37cab557de76776013928b9148f4 100644
--- a/python/paddle/fluid/tests/unittests/dist_ctr_reader.py
+++ b/python/paddle/fluid/tests/unittests/dist_ctr_reader.py
@@ -108,7 +108,7 @@ def load_lr_input_record(sent):
 feeding_index = {'dnn_input': 0, 'lr_input': 1, 'click': 2}
 
 
-class Dataset(object):
+class Dataset:
     def train(self):
         '''
         Load trainset.
diff --git a/python/paddle/fluid/tests/unittests/dist_transformer.py b/python/paddle/fluid/tests/unittests/dist_transformer.py
index 4765ee7d82bf53c09f80c9087cc9700801484b87..514fcf4b869be2e3a92cc4e6eb80bdd7d28d70b9 100644
--- a/python/paddle/fluid/tests/unittests/dist_transformer.py
+++ b/python/paddle/fluid/tests/unittests/dist_transformer.py
@@ -36,7 +36,7 @@ fluid.default_main_program().random_seed = 1
 
 
 # from transformer_config import ModelHyperParams, TrainTaskConfig, merge_cfg_from_list
-class TrainTaskConfig(object):
+class TrainTaskConfig:
     # only support GPU currently
     use_gpu = True
     # the epoch number to train.
@@ -88,7 +88,7 @@ class TrainTaskConfig(object):
     use_token_batch = False
 
 
-class InferTaskConfig(object):
+class InferTaskConfig:
     use_gpu = True
     # the number of examples in one run for sequence generation.
     batch_size = 10
@@ -105,7 +105,7 @@ class InferTaskConfig(object):
     model_path = "trained_models/pass_1.infer.model"
 
 
-class ModelHyperParams(object):
+class ModelHyperParams:
     # These following five vocabularies related configurations will be set
     # automatically according to the passed vocabulary path and special tokens.
     # size of source word dictionary.
@@ -268,7 +268,7 @@ fast_decoder_data_input_fields = (
 
 
 # from optim import LearningRateScheduler
-class LearningRateScheduler(object):
+class LearningRateScheduler:
     """
     Wrapper for learning rate scheduling as described in the Transformer paper.
     LearningRateScheduler adapts the learning rate externally and the adapted
@@ -714,13 +714,13 @@ def train_loop(
 
 
 # import transformer_reader as reader
-class SortType(object):
+class SortType:
     GLOBAL = 'global'
     POOL = 'pool'
     NONE = "none"
 
 
-class Converter(object):
+class Converter:
     def __init__(self, vocab, beg, end, unk, delimiter):
         self._vocab = vocab
         self._beg = beg
@@ -739,7 +739,7 @@ class Converter(object):
         )
 
 
-class ComposedConverter(object):
+class ComposedConverter:
     def __init__(self, converters):
         self._converters = converters
 
@@ -750,7 +750,7 @@ class ComposedConverter(object):
         ]
 
 
-class SentenceBatchCreator(object):
+class SentenceBatchCreator:
     def __init__(self, batch_size):
         self.batch = []
         self._batch_size = batch_size
@@ -763,7 +763,7 @@ class SentenceBatchCreator(object):
             return tmp
 
 
-class TokenBatchCreator(object):
+class TokenBatchCreator:
     def __init__(self, batch_size):
         self.batch = []
         self.max_len = -1
@@ -782,14 +782,14 @@ class TokenBatchCreator(object):
             self.batch.append(info)
 
 
-class SampleInfo(object):
+class SampleInfo:
     def __init__(self, i, max_len, min_len):
         self.i = i
         self.min_len = min_len
         self.max_len = max_len
 
 
-class MinMaxFilter(object):
+class MinMaxFilter:
     def __init__(self, max_len, min_len, underlying_creator):
         self._min_len = min_len
         self._max_len = max_len
@@ -806,7 +806,7 @@ class MinMaxFilter(object):
         return self._creator.batch
 
 
-class DataReader(object):
+class DataReader:
     """
     The data reader loads all data from files and produces batches of data
     in the way corresponding to settings.
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/bert_utils.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/bert_utils.py
index 0cc22162d202c10b5cc4d889a017cf2e69c0953c..79c911c82e4601b21b9488e562c65e3d305acaf8 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/bert_utils.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/bert_utils.py
@@ -223,7 +223,7 @@ def prepare_batch_data(
     return res
 
 
-class DataReader(object):
+class DataReader:
     def __init__(
         self,
         batch_size=4096,
@@ -339,7 +339,7 @@ class DataReader(object):
         return wrapper
 
 
-class ModelHyperParams(object):
+class ModelHyperParams:
     generate_neg_sample = False
     epoch = 100
     max_seq_len = 512
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py
index a931276c56e9609737754e46585ffa2ec31ce097..2604bdd3a690d42918007a936813291cc224c76c 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py
@@ -355,7 +355,7 @@ def if_with_and_or_4(x, y=None):
 
 
 def if_with_class_var(x, y=None):
-    class Foo(object):
+    class Foo:
         def __init__(self):
             self.a = 1
             self.b = 2
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/predictor_utils.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/predictor_utils.py
index 554cf95e82483ac4b8bdc56d0284f3e6a0171ded..2e49f3778ac27cce38d962103c32aeade1935a2c 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/predictor_utils.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/predictor_utils.py
@@ -21,7 +21,7 @@ from paddle.fluid.core import AnalysisConfig
 from paddle.fluid.core import create_paddle_predictor
 
 
-class PredictorTools(object):
+class PredictorTools:
     '''
     Paddle-Inference predictor
     '''
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/seq2seq_utils.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/seq2seq_utils.py
index 44a55c79c92e33ead837a5c1eae771213587596b..e93405ebd399f1d48dff187718aeea32dce66b0c 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/seq2seq_utils.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/seq2seq_utils.py
@@ -91,7 +91,7 @@ def get_data_iter(batch_size, mode='train', cache_num=20):
             yield (src_ids, src_mask, tar_ids, tar_mask)
 
 
-class Seq2SeqModelHyperParams(object):
+class Seq2SeqModelHyperParams:
     # Whether use attention model
     attention = False
 
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model.py
index 7ea2a97cb0aa3cfc2ac3d645a8dbab1a3735bf9f..0bb08405141c3fc5b328bf31d828b03e23d4f877 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model.py
@@ -21,7 +21,7 @@ from paddle.fluid.dygraph import Embedding, Layer, Linear
 from paddle.static import Variable
 
 
-class EmbeddingLayer(object):
+class EmbeddingLayer:
     """
     Embedding Layer class
     """
@@ -53,7 +53,7 @@ class EmbeddingLayer(object):
         return emb
 
 
-class FCLayer(object):
+class FCLayer:
     """
     Fully Connect Layer class
     """
@@ -79,7 +79,7 @@ class FCLayer(object):
         return fc
 
 
-class ConcatLayer(object):
+class ConcatLayer:
     """
     Connection Layer class
     """
@@ -98,7 +98,7 @@ class ConcatLayer(object):
         return concat
 
 
-class ReduceMeanLayer(object):
+class ReduceMeanLayer:
     """
     Reduce Mean Layer class
     """
@@ -117,7 +117,7 @@ class ReduceMeanLayer(object):
         return mean
 
 
-class CosSimLayer(object):
+class CosSimLayer:
     """
     Cos Similarly Calculate Layer
     """
@@ -136,7 +136,7 @@ class CosSimLayer(object):
         return sim
 
 
-class ElementwiseMaxLayer(object):
+class ElementwiseMaxLayer:
     """
     Elementwise Max Layer class
     """
@@ -155,7 +155,7 @@ class ElementwiseMaxLayer(object):
         return max
 
 
-class ElementwiseAddLayer(object):
+class ElementwiseAddLayer:
     """
     Elementwise Add Layer class
     """
@@ -174,7 +174,7 @@ class ElementwiseAddLayer(object):
         return add
 
 
-class ElementwiseSubLayer(object):
+class ElementwiseSubLayer:
     """
     Elementwise Add Layer class
     """
@@ -193,7 +193,7 @@ class ElementwiseSubLayer(object):
         return sub
 
 
-class ConstantLayer(object):
+class ConstantLayer:
     """
     Generate A Constant Layer class
     """
@@ -215,7 +215,7 @@ class ConstantLayer(object):
         return constant
 
 
-class SoftsignLayer(object):
+class SoftsignLayer:
     """
     Softsign Layer class
     """
@@ -439,7 +439,7 @@ class FC(Layer):
         return self._helper.append_activation(pre_activation, act=self._act)
 
 
-class HingeLoss(object):
+class HingeLoss:
     """
     Hing Loss Calculate class
     """
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model_v2.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model_v2.py
index ce4ba45f964e88d330040caffc678ab2a435183e..c4739b2dc37d4b38055f6adfea258a7a50b5eaee 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model_v2.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/simnet_dygraph_model_v2.py
@@ -17,7 +17,7 @@ import paddle
 from paddle.static import Variable
 
 
-class EmbeddingLayer(object):
+class EmbeddingLayer:
     """
     Embedding Layer class
     """
@@ -50,7 +50,7 @@ class EmbeddingLayer(object):
         return emb
 
 
-class FCLayer(object):
+class FCLayer:
     """
     Fully Connect Layer class
     """
@@ -76,7 +76,7 @@ class FCLayer(object):
         return fc
 
 
-class ConcatLayer(object):
+class ConcatLayer:
     """
     Connection Layer class
     """
@@ -95,7 +95,7 @@ class ConcatLayer(object):
         return concat
 
 
-class ReduceMeanLayer(object):
+class ReduceMeanLayer:
     """
     Reduce Mean Layer class
     """
@@ -114,7 +114,7 @@ class ReduceMeanLayer(object):
         return mean
 
 
-class CosSimLayer(object):
+class CosSimLayer:
     """
     Cos Similarly Calculate Layer
     """
@@ -133,7 +133,7 @@ class CosSimLayer(object):
         return sim
 
 
-class ElementwiseMaxLayer(object):
+class ElementwiseMaxLayer:
     """
     Elementwise Max Layer class
     """
@@ -152,7 +152,7 @@ class ElementwiseMaxLayer(object):
         return max
 
 
-class ElementwiseAddLayer(object):
+class ElementwiseAddLayer:
     """
     Elementwise Add Layer class
     """
@@ -171,7 +171,7 @@ class ElementwiseAddLayer(object):
         return add
 
 
-class ElementwiseSubLayer(object):
+class ElementwiseSubLayer:
     """
     Elementwise Add Layer class
     """
@@ -190,7 +190,7 @@ class ElementwiseSubLayer(object):
         return sub
 
 
-class ConstantLayer(object):
+class ConstantLayer:
     """
     Generate A Constant Layer class
     """
@@ -212,7 +212,7 @@ class ConstantLayer(object):
         return constant
 
 
-class SoftsignLayer(object):
+class SoftsignLayer:
     """
     Softsign Layer class
     """
@@ -425,7 +425,7 @@ class FC(paddle.nn.Layer):
         return self._helper.append_activation(pre_activation, act=self._act)
 
 
-class HingeLoss(object):
+class HingeLoss:
     """
     Hing Loss Calculate class
     """
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_bmn.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_bmn.py
index 7e6dad60b29941516be8f72b806ac2231902ed3c..9ce37b565b906aabd811a08d6ef31a1bd2a5a306 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_bmn.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_bmn.py
@@ -447,7 +447,7 @@ def bmn_loss_func(
     return loss, tem_loss, pem_reg_loss, pem_cls_loss
 
 
-class Args(object):
+class Args:
     epoch = 1
     batch_size = 4
     learning_rate = 0.1
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_break_continue.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_break_continue.py
index 45b149617d2bfc614d0ad9466a375c73e5e5d2f7..143b14f7ed641693d2dd1c53bb9c8de224b8e4e7 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_break_continue.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_break_continue.py
@@ -160,7 +160,7 @@ def test_for_in_else(x):
 
 
 def while_loop_class_var(x):
-    class Foo(object):
+    class Foo:
         def __init__(self):
             self.a = 3
             self.b = 4
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_cycle_gan.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_cycle_gan.py
index 756da0243b9b122309ad57a180e1c067b1e8b5b2..12a4f48f6445240d012cf2a2332fa3009cfea29b 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_cycle_gan.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_cycle_gan.py
@@ -480,7 +480,7 @@ class DeConv2D(fluid.dygraph.Layer):
         return conv
 
 
-class ImagePool(object):
+class ImagePool:
     def __init__(self, pool_size=50):
         self.pool = []
         self.count = 0
@@ -530,7 +530,7 @@ def reader_creater():
     return reader
 
 
-class Args(object):
+class Args:
     epoch = 1
     batch_size = 4
     image_shape = [3, IMAGE_SIZE, IMAGE_SIZE]
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_full_name_usage.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_full_name_usage.py
index a245b607f881dd1e3e5b5853d0e1bb1617cb95e9..f16d2410cbecc147c722b93f5138333c37366267 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_full_name_usage.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_full_name_usage.py
@@ -44,7 +44,7 @@ def decorated_call_decorated(x):
     return jit_decorated_func(x)
 
 
-class DoubleDecorated(object):
+class DoubleDecorated:
     @classmethod
     @declarative
     def double_decorated_func1(self, x):
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_lac.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_lac.py
index 3026d7e5675d8baf2b115509e86f46d7c31f2560..9d00db1caa660a8673eb842594114cede1ff36f4 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_lac.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_lac.py
@@ -472,7 +472,7 @@ class LexNet(fluid.dygraph.Layer):
         return avg_cost, crf_decode
 
 
-class Args(object):
+class Args:
     epoch = 1
     batch_size = 4
     vocab_size = 100
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_loop.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_loop.py
index 6bc237f3781b62b952e9818aa376337f698bba49..75c59897d8825de30d8ef31857173b9742294225 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_loop.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_loop.py
@@ -153,7 +153,7 @@ def while_loop_bool_op2(x):
 
 
 def while_loop_class_var(x):
-    class Foo(object):
+    class Foo:
         def __init__(self):
             self.a = 3
             self.b = 4
@@ -179,7 +179,7 @@ def loop_var_contains_property(x):
 
 
 def for_loop_class_var(max_len):
-    class Foo(object):
+    class Foo:
         def __init__(self):
             self.a = 3
             self.b = 4
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_mobile_net.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_mobile_net.py
index 2cb5451bfa8d003ac709335288c62d39d514e0ca..083345d9db99a4fbdfae516b9a41fb0660b86ec7 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_mobile_net.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_mobile_net.py
@@ -470,7 +470,7 @@ def fake_data_reader(batch_size, label_size):
     return reader
 
 
-class Args(object):
+class Args:
     batch_size = 4
     model = "MobileNetV1"
     lr = 0.001
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_reinforcement_learning.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_reinforcement_learning.py
index 9e79d40df80b23f07ceb688a0ddbdd3ad9133f72..c111e5c4820d09313a44ab1f0d6edb22d10d53e5 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_reinforcement_learning.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_reinforcement_learning.py
@@ -52,7 +52,7 @@ class Policy(Layer):
         return log_prob
 
 
-class Args(object):
+class Args:
     gamma = 0.99
     log_interval = 1
     train_step = 10
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_sentiment.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_sentiment.py
index 5c52fade779ebcb210c38f6d95e4f6dc71ddef45..25ca7e08472a0cfedea787f534c54308e3d05fa1 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_sentiment.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_sentiment.py
@@ -293,7 +293,7 @@ def fake_data_reader(class_num, vocab_size, batch_size, padding_size):
     return reader
 
 
-class Args(object):
+class Args:
     epoch = 1
     batch_size = 4
     class_num = 2
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet.py
index 466c6affcb846fcbc82422d44d822eca2b7d17c7..3e70147d302963161e28a1683eb8ee7ed40a05ed 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet.py
@@ -73,7 +73,7 @@ def fake_vocabulary():
 vocab = fake_vocabulary()
 
 
-class FakeReaderProcessor(object):
+class FakeReaderProcessor:
     def __init__(self, args, vocab):
         self.vocab = vocab
         self.seq_len = args.seq_len
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet_v2.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet_v2.py
index bbb408e48c3281fdca73cfdab24d2fdcb42e9945..b1ed858f85a125fd857a2f92b61b618b218c0096 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet_v2.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_simnet_v2.py
@@ -71,7 +71,7 @@ def fake_vocabulary():
 vocab = fake_vocabulary()
 
 
-class FakeReaderProcessor(object):
+class FakeReaderProcessor:
     def __init__(self, args, vocab):
         self.vocab = vocab
         self.seq_len = args.seq_len
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_tsm.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_tsm.py
index 5923618ca3bcf3382826b59198d090779dab3067..cc307e5a7bb16b8a80de3cadcb65d934d568344c 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_tsm.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_tsm.py
@@ -214,7 +214,7 @@ class TSM_ResNet(fluid.dygraph.Layer):
         return y
 
 
-class FakeDataReader(object):
+class FakeDataReader:
     def __init__(self, mode, cfg):
         self.format = cfg.MODEL.format
         self.num_classes = cfg.MODEL.num_classes
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_yolov3.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_yolov3.py
index eecd6806fa540b2616dd4624c3e69a3277f33f5d..086e56828a3c6b4b4be73122ee34a957b50b3ce4 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/test_yolov3.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/test_yolov3.py
@@ -29,7 +29,7 @@ random.seed(0)
 np.random.seed(0)
 
 
-class SmoothedValue(object):
+class SmoothedValue:
     """Track a series of values and provide access to smoothed values over a
     window or the global series average.
     """
@@ -46,7 +46,7 @@ class SmoothedValue(object):
         return self.loss_sum / self.iter_cnt
 
 
-class FakeDataReader(object):
+class FakeDataReader:
     def __init__(self):
         self.generator_out = []
         self.total_iter = cfg.max_iter
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_dygraph_model.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_dygraph_model.py
index b27a35ae7ec1abf283ea61032096390f601d6f09..ee32e0640ca98527abf03a50079765e20e4c939f 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_dygraph_model.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_dygraph_model.py
@@ -565,7 +565,7 @@ class WrapDecoder(Layer):
         return logits
 
 
-class CrossEntropyCriterion(object):
+class CrossEntropyCriterion:
     def __init__(self, label_smooth_eps):
         self.label_smooth_eps = label_smooth_eps
 
diff --git a/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_util.py b/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_util.py
index 7e77f1ad033db1ef6a5d18ac3db95c1421cc7395..bb0011de6bc9ed60efdcdab477cba7e079737863 100644
--- a/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_util.py
+++ b/python/paddle/fluid/tests/unittests/dygraph_to_static/transformer_util.py
@@ -97,7 +97,7 @@ fast_decoder_data_input_fields = (
 )
 
 
-class ModelHyperParams(object):
+class ModelHyperParams:
     print_step = 2
     save_dygraph_model_path = "dygraph_trained_models"
     save_static_model_path = "static_trained_models"
@@ -290,7 +290,7 @@ def get_feed_data_reader(args, mode='train'):
     return __for_train__ if mode == 'train' else __for_test__
 
 
-class InputField(object):
+class InputField:
     def __init__(self, input_slots):
         self.feed_list = []
         for slot in input_slots:
diff --git a/python/paddle/fluid/tests/unittests/feed_data_reader.py b/python/paddle/fluid/tests/unittests/feed_data_reader.py
index 9a1f15ddff5ec7dea2f1cae7b9160e705d71fec0..ef2e18a42966d840d65844b2134cd7842d410ff4 100644
--- a/python/paddle/fluid/tests/unittests/feed_data_reader.py
+++ b/python/paddle/fluid/tests/unittests/feed_data_reader.py
@@ -25,7 +25,7 @@ def cyclic_reader(reader):
     return __reader__
 
 
-class FeedDataReader(object):
+class FeedDataReader:
     def __init__(self, feed_list, reader):
         self._feed_list = []
         for var in feed_list:
diff --git a/python/paddle/fluid/tests/unittests/mlu/test_collective_api_base_mlu.py b/python/paddle/fluid/tests/unittests/mlu/test_collective_api_base_mlu.py
index 5adcf7a29875af054fad9e7b20633de9462cfae7..d15cbfc56f1dcda389ef25ab956f4bcc57da6977 100644
--- a/python/paddle/fluid/tests/unittests/mlu/test_collective_api_base_mlu.py
+++ b/python/paddle/fluid/tests/unittests/mlu/test_collective_api_base_mlu.py
@@ -39,7 +39,7 @@ def DataTypeCast(date_type):
     return np_data_type
 
 
-class TestCollectiveAPIRunnerBase(object):
+class TestCollectiveAPIRunnerBase:
     def get_model(self, train_prog, startup_prog, rank, indata=None):
         raise NotImplementedError(
             "get model should be implemented by child class."
diff --git a/python/paddle/fluid/tests/unittests/mlu/test_collective_base_mlu.py b/python/paddle/fluid/tests/unittests/mlu/test_collective_base_mlu.py
index 3e005bafb4131bc073daea78f9539a82215b9629..f4d41818b3325485a53eeb9d670fa9abf94c8a4b 100644
--- a/python/paddle/fluid/tests/unittests/mlu/test_collective_base_mlu.py
+++ b/python/paddle/fluid/tests/unittests/mlu/test_collective_base_mlu.py
@@ -51,7 +51,7 @@ def DataTypeCast(date_type):
     return np_data_type
 
 
-class TestCollectiveRunnerBase(object):
+class TestCollectiveRunnerBase:
     def get_model(self, train_prog, startup_prog, col_type):
         raise NotImplementedError(
             "get model should be implemented by child class."
diff --git a/python/paddle/fluid/tests/unittests/mlu/test_pool2d_op_mlu.py b/python/paddle/fluid/tests/unittests/mlu/test_pool2d_op_mlu.py
index 57202e62a43330917f26bdc8e4f2382f2753708c..1a7a2f2255145e0963d49c6e257c39796287f3ed 100644
--- a/python/paddle/fluid/tests/unittests/mlu/test_pool2d_op_mlu.py
+++ b/python/paddle/fluid/tests/unittests/mlu/test_pool2d_op_mlu.py
@@ -174,7 +174,7 @@ def pool2d_backward_navie(
     return x_grad
 
 
-class TestPool2D_Op_Mixin(object):
+class TestPool2D_Op_Mixin:
     def setUp(self):
         self.place = paddle.device.MLUPlace(0)
         self.__class__.use_mlu = True
diff --git a/python/paddle/fluid/tests/unittests/mlu/test_sync_batch_norm_base_mlu.py b/python/paddle/fluid/tests/unittests/mlu/test_sync_batch_norm_base_mlu.py
index cb7a890adb22271d993c05323c45ae9bd407e1e7..281d2d9de027cfa21796f271deb3ff24af89502d 100644
--- a/python/paddle/fluid/tests/unittests/mlu/test_sync_batch_norm_base_mlu.py
+++ b/python/paddle/fluid/tests/unittests/mlu/test_sync_batch_norm_base_mlu.py
@@ -40,7 +40,7 @@ paddle.enable_static()
 SEED = 10
 
 
-class TestSyncBatchNormRunnerBase(object):
+class TestSyncBatchNormRunnerBase:
     def get_model(
         self,
         main,
diff --git a/python/paddle/fluid/tests/unittests/npu/test_collective_base_npu.py b/python/paddle/fluid/tests/unittests/npu/test_collective_base_npu.py
index 55bc9dce1863ef15ffd1bd58939e38393270c8ad..3e497ced85d618f7645fce8ae782f654adb468a9 100644
--- a/python/paddle/fluid/tests/unittests/npu/test_collective_base_npu.py
+++ b/python/paddle/fluid/tests/unittests/npu/test_collective_base_npu.py
@@ -28,7 +28,7 @@ import paddle.fluid.unique_name as nameGen
 from paddle.fluid import core
 
 
-class TestCollectiveRunnerBase(object):
+class TestCollectiveRunnerBase:
     def get_model(self, train_prog, startup_prog):
         raise NotImplementedError(
             "get model should be implemented by child class."
diff --git a/python/paddle/fluid/tests/unittests/npu/test_sync_batch_norm_base_npu.py b/python/paddle/fluid/tests/unittests/npu/test_sync_batch_norm_base_npu.py
index 101a749f424913d3e57a138b6b308cb8383319e4..0003d4c44ca6f762bbea3a2383bd4b0f37d1a7c2 100644
--- a/python/paddle/fluid/tests/unittests/npu/test_sync_batch_norm_base_npu.py
+++ b/python/paddle/fluid/tests/unittests/npu/test_sync_batch_norm_base_npu.py
@@ -45,7 +45,7 @@ paddle.enable_static()
 SEED = 10
 
 
-class TestSyncBatchNormRunnerBase(object):
+class TestSyncBatchNormRunnerBase:
     def get_model(
         self,
         main,
diff --git a/python/paddle/fluid/tests/unittests/op_test.py b/python/paddle/fluid/tests/unittests/op_test.py
index baa4f26feb81c974115f21ed0912b9491b83a751..d359f3459fadb7b9aed877aeb573a72b12aeddc8 100644
--- a/python/paddle/fluid/tests/unittests/op_test.py
+++ b/python/paddle/fluid/tests/unittests/op_test.py
@@ -1566,7 +1566,7 @@ class OpTest(unittest.TestCase):
             )
             return found[0]
 
-        class Checker(object):
+        class Checker:
             """base class for check with self.outputs.
             currently don't support check between checkers.
             """
diff --git a/python/paddle/fluid/tests/unittests/ps/ps_dnn_trainer.py b/python/paddle/fluid/tests/unittests/ps/ps_dnn_trainer.py
index 1fa54bccd2414933a7548d2f495a6563257e0289..5b8c2e6e6734c5dd33ec470285a2eafb0abbe0b0 100755
--- a/python/paddle/fluid/tests/unittests/ps/ps_dnn_trainer.py
+++ b/python/paddle/fluid/tests/unittests/ps/ps_dnn_trainer.py
@@ -46,7 +46,7 @@ def is_distributed_env():
         return True
 
 
-class YamlHelper(object):
+class YamlHelper:
     def load_yaml(self, yaml_file, other_part=None):
         part_list = ["runner", "hyper_parameters"]
         if other_part:
@@ -317,7 +317,7 @@ def bf16_to_fp32(val):
     return np.float32(struct.unpack('<f', struct.pack('<I', val << 16))[0])
 
 
-class DnnTrainer(object):
+class DnnTrainer:
     def __init__(self, config):
         self.metrics = {}
         self.config = config
diff --git a/python/paddle/fluid/tests/unittests/py_precise_roi_pool.py b/python/paddle/fluid/tests/unittests/py_precise_roi_pool.py
index e18c302fd7b41a551028510db983236140ff942c..a48ed4bde0fbb74ef60a4781d5d8df771e31311a 100644
--- a/python/paddle/fluid/tests/unittests/py_precise_roi_pool.py
+++ b/python/paddle/fluid/tests/unittests/py_precise_roi_pool.py
@@ -16,7 +16,7 @@ import math
 import numpy as np
 
 
-class PyPrRoIPool(object):
+class PyPrRoIPool:
     def __init__(self):
         pass
 
diff --git a/python/paddle/fluid/tests/unittests/rnn/rnn_numpy.py b/python/paddle/fluid/tests/unittests/rnn/rnn_numpy.py
index 3b0b97010e97ab5f3f15b021af92d8728a22ba17..b589f75123d6c10cd2c6c94087478e5e0504978d 100644
--- a/python/paddle/fluid/tests/unittests/rnn/rnn_numpy.py
+++ b/python/paddle/fluid/tests/unittests/rnn/rnn_numpy.py
@@ -16,7 +16,7 @@ import numpy as np
 import math
 
 
-class LayerMixin(object):
+class LayerMixin:
     def __call__(self, *args, **kwargs):
         return self.forward(*args, **kwargs)
 
diff --git a/python/paddle/fluid/tests/unittests/spawn_runner_base.py b/python/paddle/fluid/tests/unittests/spawn_runner_base.py
index 2b35fe70cf9359d854cac9db98da2e40bed17eca..d5e0e8a20542b414ac79c64fc39f5fb407226cfb 100644
--- a/python/paddle/fluid/tests/unittests/spawn_runner_base.py
+++ b/python/paddle/fluid/tests/unittests/spawn_runner_base.py
@@ -23,7 +23,7 @@ from paddle.fluid.framework import _test_eager_guard
 
 
 # NOTE: compatible TestParallelDyGraphRunnerBase args
-class SpawnAssistTestArgs(object):
+class SpawnAssistTestArgs:
     update_method = "local"
     trainer_id = 0
     find_unused_parameters = False
diff --git a/python/paddle/fluid/tests/unittests/test_activation_op.py b/python/paddle/fluid/tests/unittests/test_activation_op.py
index a933d591a8798469ce5657b8a88b60da59c31097..ec2adcfcdc5dc4c857a29a96200c3070a5ffffb3 100755
--- a/python/paddle/fluid/tests/unittests/test_activation_op.py
+++ b/python/paddle/fluid/tests/unittests/test_activation_op.py
@@ -165,7 +165,7 @@ class TestExpm1API(unittest.TestCase):
         # The input dtype must be float16, float32, float64.
 
 
-class TestParameter(object):
+class TestParameter:
     def test_out_name(self):
         with fluid.program_guard(fluid.Program()):
             np_x = np.array([0.1])
diff --git a/python/paddle/fluid/tests/unittests/test_argsort_op.py b/python/paddle/fluid/tests/unittests/test_argsort_op.py
index d8df9fefbbd43831a4c1f7837b7826ae16637110..8dc69001ffeba7ac26f1a998c730ba8e00767ced 100644
--- a/python/paddle/fluid/tests/unittests/test_argsort_op.py
+++ b/python/paddle/fluid/tests/unittests/test_argsort_op.py
@@ -25,7 +25,7 @@ from paddle.fluid.backward import append_backward
 np.random.seed(123)
 
 
-class PyArgsort(object):
+class PyArgsort:
     def __init__(self, input_shape, axis, descending, dtype):
         self.x = np.random.random(input_shape).astype(dtype)
         self.label = np.random.random(input_shape).astype(dtype)
diff --git a/python/paddle/fluid/tests/unittests/test_auto_parallel_partitioner_gpt.py b/python/paddle/fluid/tests/unittests/test_auto_parallel_partitioner_gpt.py
index 5802e33f537277bc6275015d56ca1c03c391f456..92528009bdc35f79e02a5b0f451353a8b9379b1e 100644
--- a/python/paddle/fluid/tests/unittests/test_auto_parallel_partitioner_gpt.py
+++ b/python/paddle/fluid/tests/unittests/test_auto_parallel_partitioner_gpt.py
@@ -843,13 +843,13 @@ def gpt_pretrain_forward(train_program, startup_program):
     return train_program, startup_program, loss
 
 
-class FakeStrategy(object):
+class FakeStrategy:
     def __init__(self):
         self.amp = False
         self.recompute = False
 
 
-class FakeFleet(object):
+class FakeFleet:
     def __init__(self):
         self.user_defined_optimizer = None
         self._user_defined_strategy = FakeStrategy()
diff --git a/python/paddle/fluid/tests/unittests/test_backward.py b/python/paddle/fluid/tests/unittests/test_backward.py
index 5b4d2d9f19e16952a00d519c4d158bb44755a197..03b2c95105c83e2bf22565c414e2c09556f6fac0 100644
--- a/python/paddle/fluid/tests/unittests/test_backward.py
+++ b/python/paddle/fluid/tests/unittests/test_backward.py
@@ -20,7 +20,7 @@ import paddle
 import numpy as np
 
 
-class BackwardNet(object):
+class BackwardNet:
     """
     Abstract Base Class.
     All Net inherited this Class should implement two functions:
diff --git a/python/paddle/fluid/tests/unittests/test_chunk_eval_op.py b/python/paddle/fluid/tests/unittests/test_chunk_eval_op.py
index b82404e01a403f6f7dad4bbd16a9874ccc30a77c..d6b2bcc28c0d3134c07846d28e472e934492c67d 100644
--- a/python/paddle/fluid/tests/unittests/test_chunk_eval_op.py
+++ b/python/paddle/fluid/tests/unittests/test_chunk_eval_op.py
@@ -20,7 +20,7 @@ from paddle.fluid import Program, program_guard
 from paddle import fluid
 
 
-class Segment(object):
+class Segment:
     def __init__(self, chunk_type, start_idx, end_idx):
         self.chunk_type = chunk_type
         self.start_idx = start_idx
diff --git a/python/paddle/fluid/tests/unittests/test_collective_api_base.py b/python/paddle/fluid/tests/unittests/test_collective_api_base.py
index 4f4be4e2dcf4d5ec36af3908aeaef37a75210013..68b44a186df25d472df3e643cb00904995d72e93 100644
--- a/python/paddle/fluid/tests/unittests/test_collective_api_base.py
+++ b/python/paddle/fluid/tests/unittests/test_collective_api_base.py
@@ -99,7 +99,7 @@ def create_test_data(shape=None, dtype=None, seed=None):
         raise NotImplementedError("Unsupported dtype for creating test data.")
 
 
-class TestCollectiveAPIRunnerBase(object):
+class TestCollectiveAPIRunnerBase:
     def get_model(
         self, train_prog, startup_prog, rank, indata=None, dtype=None
     ):
diff --git a/python/paddle/fluid/tests/unittests/test_collective_base.py b/python/paddle/fluid/tests/unittests/test_collective_base.py
index 2104f1865f4a9d664959a8439136deb63f2bc779..08d6fca95a4316f6344531be5f51bfa75fd4435d 100644
--- a/python/paddle/fluid/tests/unittests/test_collective_base.py
+++ b/python/paddle/fluid/tests/unittests/test_collective_base.py
@@ -26,7 +26,7 @@ import paddle.fluid.unique_name as nameGen
 from paddle.fluid import core
 
 
-class TestCollectiveRunnerBase(object):
+class TestCollectiveRunnerBase:
     def get_model(self, train_prog, startup_prog):
         raise NotImplementedError(
             "get model should be implemented by child class."
diff --git a/python/paddle/fluid/tests/unittests/test_crf_decoding_op.py b/python/paddle/fluid/tests/unittests/test_crf_decoding_op.py
index 8336bad6de1bf447df029ba882842b62cf06220a..801a1622421cbecd20aa553ca18078186329a024 100644
--- a/python/paddle/fluid/tests/unittests/test_crf_decoding_op.py
+++ b/python/paddle/fluid/tests/unittests/test_crf_decoding_op.py
@@ -19,7 +19,7 @@ import numpy as np
 from op_test import OpTest
 
 
-class CRFDecoding(object):
+class CRFDecoding:
     def __init__(
         self, emission_weights, transition_weights, seq_start_positions
     ):
diff --git a/python/paddle/fluid/tests/unittests/test_dist_base.py b/python/paddle/fluid/tests/unittests/test_dist_base.py
index 6212de9ebcfa489c48b29eead31970a56e6ebd45..5b874f1a17b92fba717bbaa99452ffe5288ee8d3 100755
--- a/python/paddle/fluid/tests/unittests/test_dist_base.py
+++ b/python/paddle/fluid/tests/unittests/test_dist_base.py
@@ -52,7 +52,7 @@ def eprint(*args, **kwargs):
     print(*args, file=sys.stderr, **kwargs)
 
 
-class TestDistRunnerBase(object):
+class TestDistRunnerBase:
     def get_model(
         self,
         batch_size=DEFAULT_BATCH_SIZE,
@@ -586,7 +586,7 @@ class TestDistRunnerBase(object):
         print_to_out(out_losses)
 
 
-class TestParallelDyGraphRunnerBase(object):
+class TestParallelDyGraphRunnerBase:
     def get_model(self):
         raise NotImplementedError(
             "get_model should be implemented by child classes."
diff --git a/python/paddle/fluid/tests/unittests/test_dist_fleet_base.py b/python/paddle/fluid/tests/unittests/test_dist_fleet_base.py
index 6686fb03a01feac78ce1368c69f5f227f0b37904..eb21f925fa751ed2560923a6ea8a91998413bbfc 100644
--- a/python/paddle/fluid/tests/unittests/test_dist_fleet_base.py
+++ b/python/paddle/fluid/tests/unittests/test_dist_fleet_base.py
@@ -45,7 +45,7 @@ LEARNING_RATE = 0.01
 DIST_UT_PORT = 0
 
 
-class FleetDistRunnerBase(object):
+class FleetDistRunnerBase:
     """
     run_pserver,run_trainer : after init role, using transpiler split program
     net : implment by child class, the network of model
diff --git a/python/paddle/fluid/tests/unittests/test_dist_fleet_heter_base.py b/python/paddle/fluid/tests/unittests/test_dist_fleet_heter_base.py
index 18c20ef675473a98cc4a065f7af9128bbd0c5cfb..6871580b27021b8bf3ebcfa505a3a163cc66ffba 100644
--- a/python/paddle/fluid/tests/unittests/test_dist_fleet_heter_base.py
+++ b/python/paddle/fluid/tests/unittests/test_dist_fleet_heter_base.py
@@ -39,7 +39,7 @@ LEARNING_RATE = 0.01
 DIST_UT_PORT = 0
 
 
-class FleetDistHeterRunnerBase(object):
+class FleetDistHeterRunnerBase:
     """
     run_pserver,run_trainer : after init role, using transpiler split program
     net : implment by child class, the network of model
diff --git a/python/paddle/fluid/tests/unittests/test_dynrnn_gradient_check.py b/python/paddle/fluid/tests/unittests/test_dynrnn_gradient_check.py
index 8474e71e8c71a6b9cd6b43dd12005bd9c69d9af8..b6f85c9876a385776f5d76d9b9ee974ddc78ac79 100644
--- a/python/paddle/fluid/tests/unittests/test_dynrnn_gradient_check.py
+++ b/python/paddle/fluid/tests/unittests/test_dynrnn_gradient_check.py
@@ -21,7 +21,7 @@ import unittest
 from decorator_helper import prog_scope
 
 
-class Memory(object):
+class Memory:
     def __init__(self, shape, dtype='float32'):
         self.ex = np.zeros(shape=shape, dtype=dtype)
         self.cur = None
@@ -43,7 +43,7 @@ class Memory(object):
         self.cur = None
 
 
-class Output(object):
+class Output:
     def __init__(self):
         self.outs = []
 
@@ -57,7 +57,7 @@ class Output(object):
         return self.outs[-1][-1]
 
 
-class BaseRNN(object):
+class BaseRNN:
     def __init__(self, ins, mems, params, outs, num_seq=5, max_seq_len=15):
         self.num_seq = num_seq
         self.inputs = collections.defaultdict(list)
diff --git a/python/paddle/fluid/tests/unittests/test_eager_deletion_padding_rnn.py b/python/paddle/fluid/tests/unittests/test_eager_deletion_padding_rnn.py
index 622e36abd80b63faff958784ef32d8622adeb25b..8bdbe2aa9c96384fcabf455ed61820b7b12b8cb9 100644
--- a/python/paddle/fluid/tests/unittests/test_eager_deletion_padding_rnn.py
+++ b/python/paddle/fluid/tests/unittests/test_eager_deletion_padding_rnn.py
@@ -27,7 +27,7 @@ from paddle.fluid.layers.control_flow import StaticRNN as PaddingRNN
 os.environ["CPU_NUM"] = "1"
 
 
-class RNNConfig(object):
+class RNNConfig:
     def __init__(self, model_type, rnn_model):
         self.model_type = model_type
         self.rnn_model = rnn_model
@@ -93,7 +93,7 @@ class RNNConfig(object):
 
 
 # Fake data reader for test
-class Reader(object):
+class Reader:
     def get_data_iter(self, rnn_config):
         for i in range(rnn_config.max_epoch):
             x = np.zeros(
diff --git a/python/paddle/fluid/tests/unittests/test_eager_deletion_recurrent_op.py b/python/paddle/fluid/tests/unittests/test_eager_deletion_recurrent_op.py
index db9f64d1791ba87f21d05bcf422457fcc9a86618..d89aba3da840e167a173795984bb9a8294d37942 100644
--- a/python/paddle/fluid/tests/unittests/test_eager_deletion_recurrent_op.py
+++ b/python/paddle/fluid/tests/unittests/test_eager_deletion_recurrent_op.py
@@ -32,7 +32,7 @@ os.environ["CPU_NUM"] = "1"
 fluid.core._set_eager_deletion_mode(0.0, 1.0, True)
 
 
-class PyRNNBase(object):
+class PyRNNBase:
     def __init__(self, input_shape, output_shape):
         self.x = np.ones(shape=input_shape).astype("float32")
         self.y = np.zeros(shape=output_shape).astype("float32")
diff --git a/python/paddle/fluid/tests/unittests/test_faster_tokenizer_op.py b/python/paddle/fluid/tests/unittests/test_faster_tokenizer_op.py
index 37835af0fb3585dbdcabe1ce5f865ace8cb2e0a4..c231bffc9d300fa0e3607765d3a854dd13761688 100755
--- a/python/paddle/fluid/tests/unittests/test_faster_tokenizer_op.py
+++ b/python/paddle/fluid/tests/unittests/test_faster_tokenizer_op.py
@@ -127,7 +127,7 @@ class FasterTokenizer(nn.Layer):
         return input_ids, seg_ids
 
 
-class Predictor(object):
+class Predictor:
     def __init__(self, model_dir):
         model_file = os.path.join(model_dir, "inference.pdmodel")
         params_file = os.path.join(model_dir, "inference.pdiparams")
diff --git a/python/paddle/fluid/tests/unittests/test_global_var_getter_setter.py b/python/paddle/fluid/tests/unittests/test_global_var_getter_setter.py
index 9e32b43ad42f8b837d6c25f1f48bf4953561f282..19c66daade99092e04911d7e377e23f16032e95e 100644
--- a/python/paddle/fluid/tests/unittests/test_global_var_getter_setter.py
+++ b/python/paddle/fluid/tests/unittests/test_global_var_getter_setter.py
@@ -16,7 +16,7 @@ import paddle.fluid as fluid
 import unittest as unittest
 
 
-class VarInfo(object):
+class VarInfo:
     def __init__(self, var_name, var_type, writable):
         self.name = var_name
         self.type = var_type
diff --git a/python/paddle/fluid/tests/unittests/test_graph_send_ue_recv_op.py b/python/paddle/fluid/tests/unittests/test_graph_send_ue_recv_op.py
index c3bb651a3de391d3e907401d83828e56a0f8fb3e..9d4d7dc25cab7fcd2730fb2cdffce90424d36842 100644
--- a/python/paddle/fluid/tests/unittests/test_graph_send_ue_recv_op.py
+++ b/python/paddle/fluid/tests/unittests/test_graph_send_ue_recv_op.py
@@ -39,7 +39,7 @@ def get_broadcast_shape(shp1, shp2):
     return rst
 
 
-class BroadCastInfo(object):
+class BroadCastInfo:
     def __init__(self, x_shape, y_shape):
         self.x_shape = x_shape
         self.y_shape = y_shape
diff --git a/python/paddle/fluid/tests/unittests/test_hsigmoid_op.py b/python/paddle/fluid/tests/unittests/test_hsigmoid_op.py
index a905b9c2ae0246d2b61d80d31813f864fe708993..5ac218f9ebd7a2b95ac1472a4681f1ccb5902048 100644
--- a/python/paddle/fluid/tests/unittests/test_hsigmoid_op.py
+++ b/python/paddle/fluid/tests/unittests/test_hsigmoid_op.py
@@ -30,7 +30,7 @@ def find_latest_set(num):
     return 1 + int(math.floor(math.log(num, 2)))
 
 
-class CodeTable(object):
+class CodeTable:
     def __init__(self, num_classes, code):
         self.c = num_classes + code
 
@@ -44,7 +44,7 @@ class CodeTable(object):
         return self.c & (1 << bit)
 
 
-class CodeTableWithCustomTree(object):
+class CodeTableWithCustomTree:
     def __init__(self, path_table, path_code, index):
         self.ptable_ = path_table
         self.pcode_ = path_code
diff --git a/python/paddle/fluid/tests/unittests/test_imperative_ocr_attention_model.py b/python/paddle/fluid/tests/unittests/test_imperative_ocr_attention_model.py
index 531b49f23b40c7a429c0c56bb7813887f23afda7..f5a33f692bdc92088b1b76734ab10729504c5728 100644
--- a/python/paddle/fluid/tests/unittests/test_imperative_ocr_attention_model.py
+++ b/python/paddle/fluid/tests/unittests/test_imperative_ocr_attention_model.py
@@ -29,7 +29,7 @@ from test_imperative_base import new_program_scope
 from paddle.fluid.framework import _test_eager_guard
 
 
-class Config(object):
+class Config:
     '''
     config for training
     '''
diff --git a/python/paddle/fluid/tests/unittests/test_imperative_star_gan_with_gradient_penalty.py b/python/paddle/fluid/tests/unittests/test_imperative_star_gan_with_gradient_penalty.py
index c26ac7c0355de977545e9f542763a5ebf1ad1e87..96fc83402762a82884834905093fdc8664a7bc59 100644
--- a/python/paddle/fluid/tests/unittests/test_imperative_star_gan_with_gradient_penalty.py
+++ b/python/paddle/fluid/tests/unittests/test_imperative_star_gan_with_gradient_penalty.py
@@ -23,7 +23,7 @@ if fluid.is_compiled_with_cuda():
     fluid.core.globals()['FLAGS_cudnn_deterministic'] = True
 
 
-class Config(object):
+class Config:
     def __init__(self, place, sort_sum_gradient=True):
         self.place = place
 
@@ -506,7 +506,7 @@ def build_optimizer(layer, cfg, loss=None):
         return optimizer
 
 
-class DyGraphTrainModel(object):
+class DyGraphTrainModel:
     def __init__(self, cfg):
         paddle.seed(1)
         paddle.framework.random._manual_program_seed(1)
@@ -564,7 +564,7 @@ class DyGraphTrainModel(object):
         return g_loss.numpy()[0], d_loss.numpy()[0]
 
 
-class StaticGraphTrainModel(object):
+class StaticGraphTrainModel:
     def __init__(self, cfg):
         self.cfg = cfg
 
diff --git a/python/paddle/fluid/tests/unittests/test_imperative_transformer_sorted_gradient.py b/python/paddle/fluid/tests/unittests/test_imperative_transformer_sorted_gradient.py
index baf1897c861fa82be01a53edbdc741c391dbc639..1ff8ffbd8518e3bc295d22ad4a503cd586536a25 100644
--- a/python/paddle/fluid/tests/unittests/test_imperative_transformer_sorted_gradient.py
+++ b/python/paddle/fluid/tests/unittests/test_imperative_transformer_sorted_gradient.py
@@ -29,7 +29,7 @@ from utils import DyGraphProgramDescTracerTestHelper, is_equal_program
 
 
 # Copy from models
-class TrainTaskConfig(object):
+class TrainTaskConfig:
     # support both CPU and GPU now.
     use_gpu = True
     # the epoch number to train.
@@ -65,7 +65,7 @@ class TrainTaskConfig(object):
     save_freq = 10000
 
 
-class InferTaskConfig(object):
+class InferTaskConfig:
     use_gpu = True
     # the number of examples in one run for sequence generation.
     batch_size = 10
@@ -82,7 +82,7 @@ class InferTaskConfig(object):
     model_path = "trained_models/pass_1.infer.model"
 
 
-class ModelHyperParams(object):
+class ModelHyperParams:
     # These following five vocabularies related configurations will be set
     # automatically according to the passed vocabulary path and special tokens.
     # size of source word dictionary.
diff --git a/python/paddle/fluid/tests/unittests/test_inference_model_io.py b/python/paddle/fluid/tests/unittests/test_inference_model_io.py
index 4e3288fd037861b2cafc05b62ee53f7b8c650de7..1e98b7dbb258cccb658023b3ef497b7564f91512 100644
--- a/python/paddle/fluid/tests/unittests/test_inference_model_io.py
+++ b/python/paddle/fluid/tests/unittests/test_inference_model_io.py
@@ -37,7 +37,7 @@ from paddle.fluid.io import (
 paddle.enable_static()
 
 
-class InferModel(object):
+class InferModel:
     def __init__(self, list):
         self.program = list[0]
         self.feed_var_names = list[1]
diff --git a/python/paddle/fluid/tests/unittests/test_linear_chain_crf_op.py b/python/paddle/fluid/tests/unittests/test_linear_chain_crf_op.py
index 203a4ac62ca49d3650a455e5afd71be27c4e472a..d352275c263f8c6ccdaec7dce31190b7424df442 100755
--- a/python/paddle/fluid/tests/unittests/test_linear_chain_crf_op.py
+++ b/python/paddle/fluid/tests/unittests/test_linear_chain_crf_op.py
@@ -19,7 +19,7 @@ import numpy as np
 from op_test import OpTest
 
 
-class LinearChainCrfForward(object):
+class LinearChainCrfForward:
     def __init__(
         self,
         seq_start_positions,
diff --git a/python/paddle/fluid/tests/unittests/test_lr_scheduler.py b/python/paddle/fluid/tests/unittests/test_lr_scheduler.py
index 55d26bc50937907de54671a9bea2704b7652e099..d1f656a5663b69051ad81ad28f459076d70bc5e6 100644
--- a/python/paddle/fluid/tests/unittests/test_lr_scheduler.py
+++ b/python/paddle/fluid/tests/unittests/test_lr_scheduler.py
@@ -52,7 +52,7 @@ def reduce_lr_on_plateau(
     return var_list[1]
 
 
-class TestReduceOnPlateauDecay(object):
+class TestReduceOnPlateauDecay:
     def test_ReduceLR(self):
         # the decay rate must be less than 1.0
         with self.assertRaises(ValueError):
diff --git a/python/paddle/fluid/tests/unittests/test_lstm_cudnn_op.py b/python/paddle/fluid/tests/unittests/test_lstm_cudnn_op.py
index 5cac9479d91240f384b7eb37f1f5e93928f934e9..b82b38c7ba5f6401e20a13ee48baa27fb74d88ae 100644
--- a/python/paddle/fluid/tests/unittests/test_lstm_cudnn_op.py
+++ b/python/paddle/fluid/tests/unittests/test_lstm_cudnn_op.py
@@ -59,7 +59,7 @@ class RandomWeight:
 weight = RandomWeight()
 
 
-class LayerMixin(object):
+class LayerMixin:
     def __call__(self, *args, **kwargs):
         return self.forward(*args, **kwargs)
 
diff --git a/python/paddle/fluid/tests/unittests/test_matmul_op.py b/python/paddle/fluid/tests/unittests/test_matmul_op.py
index 62c7e9b7bb716f00d97d2b12e4767420bca992aa..004ac2459641c94e652640b85b350934c8524e97 100644
--- a/python/paddle/fluid/tests/unittests/test_matmul_op.py
+++ b/python/paddle/fluid/tests/unittests/test_matmul_op.py
@@ -85,7 +85,7 @@ def reference_matmul(X, Y, transpose_X=False, transpose_Y=False):
     return Out
 
 
-class Generator(object):
+class Generator:
     def setUp(self):
         self.op_type = "matmul"
         X = np.random.random(self.shape_X).astype("float32")
diff --git a/python/paddle/fluid/tests/unittests/test_matmul_op_with_head.py b/python/paddle/fluid/tests/unittests/test_matmul_op_with_head.py
index bd6b1877ecd290e0055017c7e9eb6e7f916c9696..a36d6149b221a7f5783c6976d2269c8bc392ff5f 100644
--- a/python/paddle/fluid/tests/unittests/test_matmul_op_with_head.py
+++ b/python/paddle/fluid/tests/unittests/test_matmul_op_with_head.py
@@ -111,7 +111,7 @@ def reference_matmul_mul_head(
 
 
 # Generator for multiple head
-class GeneratorMulHead(object):
+class GeneratorMulHead:
     def setUp(self):
         self.op_type = "matmul"
         X = np.random.random(self.shape_X).astype("float32")
@@ -232,7 +232,7 @@ def generate_compatible_shapes_mul_head2(
 
 
 # Generator for multiple head, case 2 when width of X is not same as height of Y
-class GeneratorMulHead2(object):
+class GeneratorMulHead2:
     def setUp(self):
         self.op_type = "matmul"
 
diff --git a/python/paddle/fluid/tests/unittests/test_optimizer_grad.py b/python/paddle/fluid/tests/unittests/test_optimizer_grad.py
index 87e39a1efb0c5aa658d7d83cab7b9e9975b55155..df2277703b2344be517bc476c796ad95d37a5ae8 100644
--- a/python/paddle/fluid/tests/unittests/test_optimizer_grad.py
+++ b/python/paddle/fluid/tests/unittests/test_optimizer_grad.py
@@ -29,7 +29,7 @@ np.random.seed(10)
 SHAPE = [16, 10]
 
 
-class SimpleNetWithCond(object):
+class SimpleNetWithCond:
     """
     Build net with conditional Block and useless layers.
     """
diff --git a/python/paddle/fluid/tests/unittests/test_paddle_multiprocessing.py b/python/paddle/fluid/tests/unittests/test_paddle_multiprocessing.py
index 4ddc3f312004ee2959c1a507e12c9cbe1dd6fe08..e050cde7ccc6903d5207ef60c5babdd79250fc69 100644
--- a/python/paddle/fluid/tests/unittests/test_paddle_multiprocessing.py
+++ b/python/paddle/fluid/tests/unittests/test_paddle_multiprocessing.py
@@ -50,7 +50,7 @@ def send_parambase(queue, event, device, dtype):
     event.wait()
 
 
-class leak_checker(object):
+class leak_checker:
     def __init__(self, test_case):
         self.checked_pids = [os.getpid()]
         self.test_case = test_case
diff --git a/python/paddle/fluid/tests/unittests/test_parallel_executor_transformer.py b/python/paddle/fluid/tests/unittests/test_parallel_executor_transformer.py
index d447860862b608ce9f6af0d40534dcb25b4daa8b..47f11e380430aa1aea5245a347dd9cfc491cc1f6 100644
--- a/python/paddle/fluid/tests/unittests/test_parallel_executor_transformer.py
+++ b/python/paddle/fluid/tests/unittests/test_parallel_executor_transformer.py
@@ -25,7 +25,7 @@ from feed_data_reader import FeedDataReader
 os.environ['CPU_NUM'] = str(4)
 
 
-class ModelHyperParams(object):
+class ModelHyperParams:
     # Dictionary size for source and target language. This model directly uses
     # paddle.dataset.wmt16 in which <bos>, <eos> and <unk> token has
     # alreay been added, but the <pad> token is not added. Transformer requires
diff --git a/python/paddle/fluid/tests/unittests/test_pool2d_op.py b/python/paddle/fluid/tests/unittests/test_pool2d_op.py
index b2ae6318cc5deb28d39f79e78c06f016c4243ef8..3692ef86279ef08f7f5792f4301716293971b295 100644
--- a/python/paddle/fluid/tests/unittests/test_pool2d_op.py
+++ b/python/paddle/fluid/tests/unittests/test_pool2d_op.py
@@ -286,7 +286,7 @@ def pool2D_forward_naive(
     return out
 
 
-class TestPool2D_Op_Mixin(object):
+class TestPool2D_Op_Mixin:
     def setUp(self):
         self.op_type = "pool2d"
         self.use_cudnn = False
diff --git a/python/paddle/fluid/tests/unittests/test_recurrent_op.py b/python/paddle/fluid/tests/unittests/test_recurrent_op.py
index 67149212a8ce6a7e364168f9aa1126ddad3d8402..29c12daf55ff3d7b0496273ce057758f7259421f 100644
--- a/python/paddle/fluid/tests/unittests/test_recurrent_op.py
+++ b/python/paddle/fluid/tests/unittests/test_recurrent_op.py
@@ -27,7 +27,7 @@ from paddle.fluid.backward import append_backward
 np.random.seed(123)
 
 
-class PyRNNBase(object):
+class PyRNNBase:
     def __init__(self, input_shape, output_shape):
         self.x = np.ones(shape=input_shape).astype("float32")
         self.y = np.zeros(shape=output_shape).astype("float32")
diff --git a/python/paddle/fluid/tests/unittests/test_rnn_decode_api.py b/python/paddle/fluid/tests/unittests/test_rnn_decode_api.py
index 00bf9735fa513616e5d7548164a475b7299d3a27..41e894badd164fb4bef3f53e8b0013a73515e576 100644
--- a/python/paddle/fluid/tests/unittests/test_rnn_decode_api.py
+++ b/python/paddle/fluid/tests/unittests/test_rnn_decode_api.py
@@ -105,7 +105,7 @@ class DecoderCell(layers.RNNCell):
         return out, [new_lstm_states, out]
 
 
-class Encoder(object):
+class Encoder:
     def __init__(self, num_layers, hidden_size, dropout_prob=0.0):
         self.encoder_cell = EncoderCell(num_layers, hidden_size, dropout_prob)
 
@@ -119,7 +119,7 @@ class Encoder(object):
         return encoder_output, encoder_final_state
 
 
-class Decoder(object):
+class Decoder:
     def __init__(
         self,
         num_layers,
@@ -191,7 +191,7 @@ class Decoder(object):
         return decoder_output, decoder_final_state, dec_seq_lengths
 
 
-class Seq2SeqModel(object):
+class Seq2SeqModel:
     """Seq2Seq model: RNN encoder-decoder with attention"""
 
     def __init__(
@@ -302,7 +302,7 @@ class Seq2SeqModel(object):
         return probs, samples, sample_length
 
 
-class PolicyGradient(object):
+class PolicyGradient:
     """policy gradient"""
 
     def __init__(self, lr=None):
@@ -395,7 +395,7 @@ def reward_func(samples, sample_length):
     )
 
 
-class MLE(object):
+class MLE:
     """teacher-forcing MLE training"""
 
     def __init__(self, lr=None):
@@ -413,7 +413,7 @@ class MLE(object):
         return loss
 
 
-class SeqPGAgent(object):
+class SeqPGAgent:
     def __init__(
         self,
         model_cls,
diff --git a/python/paddle/fluid/tests/unittests/test_viterbi_decode_op.py b/python/paddle/fluid/tests/unittests/test_viterbi_decode_op.py
index 624f2f5e6165b646d19f83a8ef3bc3b979b34f6d..b86be49eaed9dc4b1502ed7de61b9eb9cc020482 100644
--- a/python/paddle/fluid/tests/unittests/test_viterbi_decode_op.py
+++ b/python/paddle/fluid/tests/unittests/test_viterbi_decode_op.py
@@ -18,7 +18,7 @@ import paddle
 paddle.enable_static()
 
 
-class Decoder(object):
+class Decoder:
     def __init__(self, transitions, use_tag=True):
         self.transitions = transitions
         self.use_tag = use_tag
diff --git a/python/paddle/fluid/tests/unittests/test_warpctc_op.py b/python/paddle/fluid/tests/unittests/test_warpctc_op.py
index 91bebab2f6c346198c995990e05d26b46ce2106e..b3febb9b40daa1cf75deb90196d0caee90a0779b 100644
--- a/python/paddle/fluid/tests/unittests/test_warpctc_op.py
+++ b/python/paddle/fluid/tests/unittests/test_warpctc_op.py
@@ -28,7 +28,7 @@ paddle.enable_static()
 CUDA_BLOCK_SIZE = 32
 
 
-class CTCForward(object):
+class CTCForward:
     def __init__(
         self,
         softmax,
diff --git a/python/paddle/fluid/tests/unittests/tokenizer/bert_tokenizer.py b/python/paddle/fluid/tests/unittests/tokenizer/bert_tokenizer.py
index cd3546ebe3cd07fe6904e10acfc8bef1be7f0e50..59c530b7d8aa932dd864fc8d55a46c8ff4d0f2a5 100755
--- a/python/paddle/fluid/tests/unittests/tokenizer/bert_tokenizer.py
+++ b/python/paddle/fluid/tests/unittests/tokenizer/bert_tokenizer.py
@@ -26,7 +26,7 @@ from tokenizer_utils import (
 )
 
 
-class BasicTokenizer(object):
+class BasicTokenizer:
     """
     Runs basic tokenization (punctuation splitting, lower casing, etc.).
     Args:
@@ -165,7 +165,7 @@ class BasicTokenizer(object):
         return "".join(output)
 
 
-class WordpieceTokenizer(object):
+class WordpieceTokenizer:
     """
     Runs WordPiece tokenization.
     Args:
diff --git a/python/paddle/fluid/tests/unittests/tokenizer/tokenizer_utils.py b/python/paddle/fluid/tests/unittests/tokenizer/tokenizer_utils.py
index 9e60b29ffb7abf34d5a1cd30b5ff5aee75cdfa35..2280292670316b46bee91fe8a9f34758197b8edf 100644
--- a/python/paddle/fluid/tests/unittests/tokenizer/tokenizer_utils.py
+++ b/python/paddle/fluid/tests/unittests/tokenizer/tokenizer_utils.py
@@ -147,7 +147,7 @@ def tokenize_chinese_chars(text):
     return output
 
 
-class PretrainedTokenizer(object):
+class PretrainedTokenizer:
     """
     The base class for all pretrained tokenizers. It mainly provides common methods
     for loading (construction and loading) and saving pretrained tokenizers. Loading
diff --git a/python/paddle/fluid/tests/unittests/utils.py b/python/paddle/fluid/tests/unittests/utils.py
index 03993c2355eea3931d8333c14f581f47f03683e4..04c3085c3dffc105e982856ecd382c5c520e4d99 100644
--- a/python/paddle/fluid/tests/unittests/utils.py
+++ b/python/paddle/fluid/tests/unittests/utils.py
@@ -101,7 +101,7 @@ def load_dygraph_vars_to_scope(model_path, scope, place):
         load_dict_to_scope(scope, opti_dict)
 
 
-class DyGraphProgramDescTracerTestHelper(object):
+class DyGraphProgramDescTracerTestHelper:
     def __init__(self, unittest_obj):
         self.unittest_obj = unittest_obj
 
diff --git a/python/paddle/fluid/tests/unittests/xpu/get_test_cover_info.py b/python/paddle/fluid/tests/unittests/xpu/get_test_cover_info.py
index f1276f765a87e644f029a90c2d473d565f127c97..d645462c7dc98b5e464d37060aaecc94def68a80 100644
--- a/python/paddle/fluid/tests/unittests/xpu/get_test_cover_info.py
+++ b/python/paddle/fluid/tests/unittests/xpu/get_test_cover_info.py
@@ -98,7 +98,7 @@ xpu_test_device_op_white_list = []
 xpu_test_device_op_type_white_list = []
 
 
-class XPUOpTestWrapper(object):
+class XPUOpTestWrapper:
     def create_classes(self):
         base_class = None
         classes = []
diff --git a/python/paddle/fluid/tests/unittests/xpu/test_collective_base_xpu.py b/python/paddle/fluid/tests/unittests/xpu/test_collective_base_xpu.py
index 1c00f3b94b8605469cebceed67ad9a2298a75225..b9be6077d4b4ffdae4e628cb4f9bf30ff95a9466 100644
--- a/python/paddle/fluid/tests/unittests/xpu/test_collective_base_xpu.py
+++ b/python/paddle/fluid/tests/unittests/xpu/test_collective_base_xpu.py
@@ -49,7 +49,7 @@ def DataTypeCast(date_type):
     return np_data_type
 
 
-class TestCollectiveRunnerBase(object):
+class TestCollectiveRunnerBase:
     def get_model(self, train_prog, startup_prog):
         raise NotImplementedError(
             "get model should be implemented by child class."
diff --git a/python/paddle/fluid/trainer_desc.py b/python/paddle/fluid/trainer_desc.py
index 484042c53a8ce866103b72de75ff3533fe8a588f..96204a2399df63b5b6ee53799c06ee04f4dbf0ab 100644
--- a/python/paddle/fluid/trainer_desc.py
+++ b/python/paddle/fluid/trainer_desc.py
@@ -26,7 +26,7 @@ __all__ = [
 ]
 
 
-class TrainerDesc(object):
+class TrainerDesc:
     '''
     Set proto from python to c++.
     Can be initialized from train_desc.
diff --git a/python/paddle/fluid/trainer_factory.py b/python/paddle/fluid/trainer_factory.py
index 761895fe3044fe5684240338c341f1622a2f6a7c..281fbd8693c2d38f4337b01bbb7cc52dec23772f 100644
--- a/python/paddle/fluid/trainer_factory.py
+++ b/python/paddle/fluid/trainer_factory.py
@@ -45,7 +45,7 @@ from multiprocessing import Process, Manager
 __all__ = ["TrainerFactory", "FetchHandlerMonitor"]
 
 
-class TrainerFactory(object):
+class TrainerFactory:
     """
     Create trainer and device worker.
     If opt_info is not None, it will get configs from opt_info,
@@ -151,7 +151,7 @@ class TrainerFactory(object):
         return trainer
 
 
-class FetchHandlerMonitor(object):
+class FetchHandlerMonitor:
     """
     Defination of FetchHandlerMonitor class,
     it's for fetch handler.
diff --git a/python/paddle/fluid/transpiler/collective.py b/python/paddle/fluid/transpiler/collective.py
index aedbaeb43cde7274a1adf9f1bdb292a5d7564a89..c22b1746966f141c2eeba08e56e01b0d641f4ddc 100644
--- a/python/paddle/fluid/transpiler/collective.py
+++ b/python/paddle/fluid/transpiler/collective.py
@@ -31,7 +31,7 @@ __all__ = ['GradAllReduce', 'LocalSGD', 'MultiThread']
 OpRole = core.op_proto_and_checker_maker.OpRole
 
 
-class Collective(object):
+class Collective:
     ''' '''
 
     def __init__(self, nrings):
diff --git a/python/paddle/fluid/transpiler/details/ufind.py b/python/paddle/fluid/transpiler/details/ufind.py
index fef6f24570c17bfc28dc87d699891d84c292a59e..f6a05cbf86c6b938e83dd4d4f7c7476fdcb5c1e6 100644
--- a/python/paddle/fluid/transpiler/details/ufind.py
+++ b/python/paddle/fluid/transpiler/details/ufind.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class UnionFind(object):
+class UnionFind:
     """Union-find data structure.
 
     Union-find is a data structure that keeps track of a set of elements partitioned
diff --git a/python/paddle/fluid/transpiler/details/vars_distributed.py b/python/paddle/fluid/transpiler/details/vars_distributed.py
index 32a27ed33b0b013166c8810aced35f8779a4e9be..cdd488bfa1d6e00cb4cc01b20047d95dfb1301ec 100644
--- a/python/paddle/fluid/transpiler/details/vars_distributed.py
+++ b/python/paddle/fluid/transpiler/details/vars_distributed.py
@@ -14,7 +14,7 @@
 from paddle.fluid.framework import Variable
 
 
-class VarStruct(object):
+class VarStruct:
     """
     record part properties of a Variable in python.
     """
@@ -28,7 +28,7 @@ class VarStruct(object):
         self.persistable = persistable
 
 
-class VarDistributed(object):
+class VarDistributed:
     """
     a class to record the var distributed on parameter servers.
     the class will record the relationship between origin var and slice var.
@@ -146,7 +146,7 @@ class VarDistributed(object):
         )
 
 
-class VarsDistributed(object):
+class VarsDistributed:
     """
     a gather about VarDistributed with many methods to find distributed vars.
     through the class, we can get overview about the distributed parameters on parameter servers.
diff --git a/python/paddle/fluid/transpiler/distribute_transpiler.py b/python/paddle/fluid/transpiler/distribute_transpiler.py
index 6bdddddd93fe36b24c5d40208ce012e4030f5426..ee98dc94a6f1ef5b036b7203a9e50bb8785c7c3c 100644
--- a/python/paddle/fluid/transpiler/distribute_transpiler.py
+++ b/python/paddle/fluid/transpiler/distribute_transpiler.py
@@ -143,7 +143,7 @@ def slice_variable(var_list, slice_count, min_block_size):
     return blocks
 
 
-class DistributeTranspilerConfig(object):
+class DistributeTranspilerConfig:
     """
         :api_attr: Static Graph
 
@@ -248,7 +248,7 @@ class DistributeTranspilerConfig(object):
         self.__sync_mode = value
 
 
-class ServerRuntimeConfig(object):
+class ServerRuntimeConfig:
     def __init__(self):
         self._rpc_send_thread_num = int(
             os.getenv("FLAGS_rpc_send_thread_num", "12")
@@ -261,7 +261,7 @@ class ServerRuntimeConfig(object):
         )
 
 
-class DistributeTranspiler(object):
+class DistributeTranspiler:
     """
         :api_attr: Static Graph
 
diff --git a/python/paddle/fluid/transpiler/ps_dispatcher.py b/python/paddle/fluid/transpiler/ps_dispatcher.py
index 589f2cb26e22f3bd2be92d8732ac82410c2b3a74..c42472f5a15c57957decff2d59ab67e7b36d7e12 100644
--- a/python/paddle/fluid/transpiler/ps_dispatcher.py
+++ b/python/paddle/fluid/transpiler/ps_dispatcher.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class PSDispatcher(object):
+class PSDispatcher:
     """
     PSDispatcher is the base class for dispatching vars
     into different pserver instance.
diff --git a/python/paddle/fluid/unique_name.py b/python/paddle/fluid/unique_name.py
index 1c0c28dae8bda2e6e1c8d0d5336c7f3c820f7e45..c881196ef56a348345c7776977225cc0bc14deae 100644
--- a/python/paddle/fluid/unique_name.py
+++ b/python/paddle/fluid/unique_name.py
@@ -18,7 +18,7 @@ from .wrapped_decorator import signature_safe_contextmanager
 __all__ = ['generate', 'switch', 'guard']
 
 
-class UniqueNameGenerator(object):
+class UniqueNameGenerator:
     """
     Generate unique name with prefix.
 
@@ -47,7 +47,7 @@ class UniqueNameGenerator(object):
         return self.prefix + "_".join([key, str(tmp)])
 
 
-class DygraphParameterNameChecker(object):
+class DygraphParameterNameChecker:
     """
     Check whether the name of parameter is used.
     """
diff --git a/python/paddle/hapi/callbacks.py b/python/paddle/hapi/callbacks.py
index c4ccb0341ebd7fb4cb543693b7f114f6f505aa45..f9d280d76dc54eb3dd89888ea06a7400436ca735 100644
--- a/python/paddle/hapi/callbacks.py
+++ b/python/paddle/hapi/callbacks.py
@@ -69,7 +69,7 @@ def config_callbacks(
     return cbk_list
 
 
-class CallbackList(object):
+class CallbackList:
     def __init__(self, callbacks=None):
         # copy
         self.callbacks = [c for c in callbacks]
@@ -129,7 +129,7 @@ class CallbackList(object):
         self._call(name, step, logs)
 
 
-class Callback(object):
+class Callback:
     """
     Base class used to build new callbacks. And new callbacks could also
     terminate training by setting `model.stop_training=True`.
diff --git a/python/paddle/hapi/model.py b/python/paddle/hapi/model.py
index 257024d048ae1e3ff16e49f3a2aecbfcbb31c86b..b7813932d86f662fa19da0d802c537392f5d2f35 100644
--- a/python/paddle/hapi/model.py
+++ b/python/paddle/hapi/model.py
@@ -259,7 +259,7 @@ def _update_input_info(inputs):
     return shapes, dtypes
 
 
-class StaticGraphAdapter(object):
+class StaticGraphAdapter:
     """
     Model traning/inference with a static graph.
     """
@@ -734,7 +734,7 @@ class StaticGraphAdapter(object):
         self._compiled_progs[mode] = compiled_prog
 
 
-class DynamicGraphAdapter(object):
+class DynamicGraphAdapter:
     def __init__(self, model):
         super().__init__()
         self.model = model
@@ -1006,7 +1006,7 @@ class DynamicGraphAdapter(object):
             self.model._scaler = None
 
 
-class Model(object):
+class Model:
     """
     An Model object is network with training and inference features.
     Dynamic graph and static graph are supported at the same time,
diff --git a/python/paddle/hapi/progressbar.py b/python/paddle/hapi/progressbar.py
index 43b913a6fc120da524288d44562b8423c2c85ac0..77f090471dfb91b788c50b2d6da13b25e22056d6 100644
--- a/python/paddle/hapi/progressbar.py
+++ b/python/paddle/hapi/progressbar.py
@@ -22,7 +22,7 @@ from collections import namedtuple
 __all__ = []
 
 
-class ProgressBar(object):
+class ProgressBar:
     """progress bar"""
 
     def __init__(
diff --git a/python/paddle/hapi/static_flops.py b/python/paddle/hapi/static_flops.py
index a3b787962fcc4cdb950d2ada35db735eb21f7edc..c72bcc4273c52818935c027dc652c728ac24b547 100644
--- a/python/paddle/hapi/static_flops.py
+++ b/python/paddle/hapi/static_flops.py
@@ -19,7 +19,7 @@ from paddle.static import Program, Variable
 __all__ = []
 
 
-class VarWrapper(object):
+class VarWrapper:
     def __init__(self, var, graph):
         assert isinstance(var, Variable)
         assert isinstance(graph, GraphWrapper)
@@ -39,7 +39,7 @@ class VarWrapper(object):
         return self._var.shape
 
 
-class OpWrapper(object):
+class OpWrapper:
     def __init__(self, op, graph):
         assert isinstance(graph, GraphWrapper)
         self._op = op
@@ -69,7 +69,7 @@ class OpWrapper(object):
         return [self._graph.var(var_name) for var_name in self._op.output(name)]
 
 
-class GraphWrapper(object):
+class GraphWrapper:
     """
     It is a wrapper of paddle.fluid.framework.IrGraph with some special functions
     for paddle slim framework.
@@ -208,7 +208,7 @@ def static_flops(program, print_detail=False):
     return _graph_flops(graph, detail=print_detail)
 
 
-class Table(object):
+class Table:
     def __init__(self, table_heads):
         self.table_heads = table_heads
         self.table_len = []
diff --git a/python/paddle/incubate/autograd/functional.py b/python/paddle/incubate/autograd/functional.py
index 6f34be56e71fe19e72dba8a26291242c407234b8..218c125a92e1a07717bdc8881ff6add051cc6ad0 100644
--- a/python/paddle/incubate/autograd/functional.py
+++ b/python/paddle/incubate/autograd/functional.py
@@ -168,7 +168,7 @@ def _zeros_like_with_grad(xs):
     return ys
 
 
-class Jacobian(object):
+class Jacobian:
     r"""
     Computes the Jacobian matrix of a given function.
 
@@ -257,7 +257,7 @@ class Jacobian(object):
         return self._jacobian.shape
 
 
-class Hessian(object):
+class Hessian:
     """
     Computes the Hessian matrix  with a given ``func`` with respect to ``xs`` .
 
@@ -328,7 +328,7 @@ class Hessian(object):
         return self.symbolic.shape
 
 
-class _Jacobian(object):
+class _Jacobian:
     """The base class for computing Jacobian matrix.
 
     ``_Jacobian`` implementes the core logic of multidimensional index and lazy
diff --git a/python/paddle/incubate/autograd/primreg.py b/python/paddle/incubate/autograd/primreg.py
index 7d81847f9052c640379dff08f31464dda165a446..cce8c49eb4f51facc7e3625b05d32efc88b8cebd 100644
--- a/python/paddle/incubate/autograd/primreg.py
+++ b/python/paddle/incubate/autograd/primreg.py
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 
-class Registry(object):
+class Registry:
     """A general registry object."""
 
     __slots__ = ['name', 'tab']
diff --git a/python/paddle/incubate/autograd/primx.py b/python/paddle/incubate/autograd/primx.py
index cdf16e77b1fbc2b60e95a75d94b16cdfae2eb6af..601b486d354c790b5e1cd6d1d23750fe098f87eb 100644
--- a/python/paddle/incubate/autograd/primx.py
+++ b/python/paddle/incubate/autograd/primx.py
@@ -115,7 +115,7 @@ def output_vars_on_path(path):
     return vars
 
 
-class VarMap(object):
+class VarMap:
     """A general map data structure for linking variables to variables.
 
     An example is linking variables to their gradients.
@@ -180,7 +180,7 @@ class VarMap(object):
 
 
 # TODO(lml): supporting control flow, nested blocks, and block other than current block of main program.
-class Transform(object):
+class Transform:
     """An object that maintains the state of transformations applied to a
     primitve program."""
 
diff --git a/python/paddle/incubate/autograd/utils.py b/python/paddle/incubate/autograd/utils.py
index 2b8082bf48de7a83b585e38418ca8951fab548ae..5437401aecaab96495961c61f7006cb9b18933b7 100644
--- a/python/paddle/incubate/autograd/utils.py
+++ b/python/paddle/incubate/autograd/utils.py
@@ -17,7 +17,7 @@ import paddle
 from paddle.fluid import framework as framework
 
 
-class PrimOption(object):
+class PrimOption:
     def __init__(self):
         self.enable_prim = False
 
diff --git a/python/paddle/jit/layer.py b/python/paddle/jit/layer.py
index 25d9ca5dbccebe7486e1759905aa48021a646a80..e3204ab65df022ce1527366e8d22b40244746153 100644
--- a/python/paddle/jit/layer.py
+++ b/python/paddle/jit/layer.py
@@ -17,7 +17,7 @@ from paddle.fluid import core
 from paddle.fluid.core import Load
 
 
-class Layer(object):
+class Layer:
     def __init__(self):
         self.cpp_layer = None
         # {name: Function}
diff --git a/python/paddle/nn/layer/rnn.py b/python/paddle/nn/layer/rnn.py
index bbada10aaf0a2523f918e8f46a23c9a304b9ba68..dba69b9848a289c54df74cc6b563f9df18afbb5f 100644
--- a/python/paddle/nn/layer/rnn.py
+++ b/python/paddle/nn/layer/rnn.py
@@ -188,7 +188,7 @@ class RNNCellBase(Layer):
                 return True
             return isinstance(seq, Sequence) and not isinstance(seq, str)
 
-        class Shape(object):
+        class Shape:
             def __init__(self, shape):
                 self.shape = shape if shape[0] == -1 else ([-1] + list(shape))
 
diff --git a/python/paddle/nn/utils/spectral_norm_hook.py b/python/paddle/nn/utils/spectral_norm_hook.py
index 288e5ea59c91a8f841d39f24e2f10600365202cb..f035d0b443116059f726f6c6d5fa7f1ca79e0a22 100644
--- a/python/paddle/nn/utils/spectral_norm_hook.py
+++ b/python/paddle/nn/utils/spectral_norm_hook.py
@@ -26,7 +26,7 @@ def normal_(x, mean=0.0, std=1.0):
     return x
 
 
-class SpectralNorm(object):
+class SpectralNorm:
     def __init__(self, name='weight', n_power_iterations=1, dim=0, eps=1e-12):
         self.name = name
         self.dim = dim
diff --git a/python/paddle/nn/utils/weight_norm_hook.py b/python/paddle/nn/utils/weight_norm_hook.py
index cdcb97aa9a0dc25b6b15e9c46463ffccf63e6e4d..2a4d07647929b41624de85bbb82e9e29b8233417 100755
--- a/python/paddle/nn/utils/weight_norm_hook.py
+++ b/python/paddle/nn/utils/weight_norm_hook.py
@@ -95,7 +95,7 @@ def _weight_norm(v, g, dim):
     return weight
 
 
-class WeightNorm(object):
+class WeightNorm:
     def __init__(self, name, dim):
         if dim is None:
             dim = -1
diff --git a/python/paddle/optimizer/lr.py b/python/paddle/optimizer/lr.py
index 3309bb4f8c7817fea737599a7a893c6e4bff1274..6f96e12f995d8852850b1a4e94bf1d522973409a 100644
--- a/python/paddle/optimizer/lr.py
+++ b/python/paddle/optimizer/lr.py
@@ -39,7 +39,7 @@ __all__ = [  # noqa
 ]
 
 
-class LRScheduler(object):
+class LRScheduler:
     """
 
     LRScheduler Base class. Define the common interface of a learning rate scheduler.
diff --git a/python/paddle/optimizer/optimizer.py b/python/paddle/optimizer/optimizer.py
index 783b11b90864137d3b76f602125a90c8fceaa40f..36aa9c151d7d3b7ebc96878b0e860647f3650bbe 100644
--- a/python/paddle/optimizer/optimizer.py
+++ b/python/paddle/optimizer/optimizer.py
@@ -98,7 +98,7 @@ def append_backward_new(
     return params_and_grads
 
 
-class Optimizer(object):
+class Optimizer:
     r"""Optimizer Base class.
 
     Define the common interface of an optimizer.
diff --git a/python/paddle/profiler/timer.py b/python/paddle/profiler/timer.py
index 311fc373c5f7ded55d188343ed181ce94adf127d..8bb49a2f37e8ee999409b708e4ff8bd2b48c70f9 100644
--- a/python/paddle/profiler/timer.py
+++ b/python/paddle/profiler/timer.py
@@ -16,7 +16,7 @@ import timeit
 from collections import OrderedDict
 
 
-class Stack(object):
+class Stack:
     """
     The stack in a Last-In/First-Out (LIFO) manner. New element is added at
     the end and an element is removed from that end.
@@ -41,7 +41,7 @@ class Stack(object):
             return None
 
 
-class Event(object):
+class Event:
     """
     A Event is used to record the cost of every step and the cost of
     the total steps except skipped steps.
@@ -152,7 +152,7 @@ class Event(object):
         return summary
 
 
-class Hook(object):
+class Hook:
     """
     As the base class. All types of hooks should inherit from it.
     """
@@ -299,7 +299,7 @@ class TimerHook(Hook):
         )
 
 
-class TimeAverager(object):
+class TimeAverager:
     """
     Record the cost of every step and count the average.
     """
@@ -346,7 +346,7 @@ class TimeAverager(object):
         return float(self._total_iters) / self._total_time
 
 
-class Benchmark(object):
+class Benchmark:
     """
     A tool for the statistics of model performance. The `before_reader`
     and `after_reader` are called in the DataLoader to count the cost
diff --git a/python/paddle/static/input.py b/python/paddle/static/input.py
index 62083769e56ca6af267832e5c65007c7d75267ee..f6e979dbcbf72f94ffa346054c33f9271fced3e8 100644
--- a/python/paddle/static/input.py
+++ b/python/paddle/static/input.py
@@ -120,7 +120,7 @@ def data(name, shape, dtype=None, lod_level=0):
         )
 
 
-class InputSpec(object):
+class InputSpec:
     """
     InputSpec describes the signature information of the model input, such as ``shape`` , ``dtype`` , ``name`` .
 
diff --git a/python/paddle/tensor/to_string.py b/python/paddle/tensor/to_string.py
index 4739628f7176b4023abb618759dce57d8d929f98..bb38d152b56082181b27ffa9277b9f19f47812fe 100644
--- a/python/paddle/tensor/to_string.py
+++ b/python/paddle/tensor/to_string.py
@@ -19,7 +19,7 @@ from paddle.fluid.data_feeder import check_type, convert_dtype
 __all__ = []
 
 
-class PrintOptions(object):
+class PrintOptions:
     precision = 8
     threshold = 1000
     edgeitems = 3
diff --git a/python/paddle/text/datasets/movielens.py b/python/paddle/text/datasets/movielens.py
index b01d8e94bb3194701eb9dd2368c9550e6a437a69..c4d0681f42a4ce8f412a9f623451926d701b9366 100644
--- a/python/paddle/text/datasets/movielens.py
+++ b/python/paddle/text/datasets/movielens.py
@@ -27,7 +27,7 @@ URL = 'https://dataset.bj.bcebos.com/movielens%2Fml-1m.zip'
 MD5 = 'c4d9eecfca2ab87c1945afe126590906'
 
 
-class MovieInfo(object):
+class MovieInfo:
     """
     Movie id, title and categories information are stored in MovieInfo.
     """
@@ -58,7 +58,7 @@ class MovieInfo(object):
         return self.__str__()
 
 
-class UserInfo(object):
+class UserInfo:
     """
     User id, gender, age, and job information are stored in UserInfo.
     """
diff --git a/python/paddle/utils/cpp_extension/cpp_extension.py b/python/paddle/utils/cpp_extension/cpp_extension.py
index 3cbfb6d732fd8ec2b015a07f57ccc4d3f60898a0..c05be5f2a4947257efad46d71411e0d28d0064c4 100644
--- a/python/paddle/utils/cpp_extension/cpp_extension.py
+++ b/python/paddle/utils/cpp_extension/cpp_extension.py
@@ -353,7 +353,7 @@ def _generate_extension_name(sources):
     return '_'.join(file_prefix)
 
 
-class BuildExtension(build_ext, object):
+class BuildExtension(build_ext):
     """
     Inherited from setuptools.command.build_ext to customize how to apply
     compilation process with share library.
@@ -724,7 +724,7 @@ class BuildExtension(build_ext, object):
                 )
 
 
-class EasyInstallCommand(easy_install, object):
+class EasyInstallCommand(easy_install):
     """
     Extend easy_intall Command to control the behavior of naming shared library
     file.
@@ -759,7 +759,7 @@ class EasyInstallCommand(easy_install, object):
                 assert os.path.exists(new_so_path)
 
 
-class BuildCommand(build, object):
+class BuildCommand(build):
     """
     Extend build Command to control the behavior of specifying `build_base` root directory.
 
diff --git a/python/paddle/utils/download.py b/python/paddle/utils/download.py
index a77d7b60dea5074d166302f50b6ac4c0a2d157d2..660e09e8668d5de2b398749872a40d70200b3037 100644
--- a/python/paddle/utils/download.py
+++ b/python/paddle/utils/download.py
@@ -27,7 +27,7 @@ try:
     from tqdm import tqdm
 except:
 
-    class tqdm(object):
+    class tqdm:
         def __init__(self, total=None):
             self.total = total
             self.n = 0
diff --git a/python/paddle/utils/op_version.py b/python/paddle/utils/op_version.py
index 9f9ae4d73c7bb0f0d223e3f8ee7ed96686b0d4a6..793e0b621990c35c5e6cb993bfb50ec6ed7a59e4 100644
--- a/python/paddle/utils/op_version.py
+++ b/python/paddle/utils/op_version.py
@@ -28,7 +28,7 @@ def Singleton(cls):
     return _singleton
 
 
-class OpUpdateInfoHelper(object):
+class OpUpdateInfoHelper:
     def __init__(self, info):
         self._info = info
 
@@ -47,7 +47,7 @@ class OpUpdateInfoHelper(object):
 
 
 @Singleton
-class OpLastCheckpointChecker(object):
+class OpLastCheckpointChecker:
     def __init__(self):
         self.raw_version_map = core.get_op_version_map()
         self.checkpoints_map = {}
diff --git a/python/paddle/utils/profiler.py b/python/paddle/utils/profiler.py
index 27803cfa442dd6effc0afe2a7121c2ed780b9200..625900e87e86849aacf41f876919bc444fc6541a 100644
--- a/python/paddle/utils/profiler.py
+++ b/python/paddle/utils/profiler.py
@@ -34,7 +34,7 @@ __all__ = [  # noqa
 ]
 
 
-class ProfilerOptions(object):
+class ProfilerOptions:
     def __init__(self, options=None):
         self.options = {
             'state': 'All',
@@ -74,7 +74,7 @@ class ProfilerOptions(object):
 _current_profiler = None
 
 
-class Profiler(object):
+class Profiler:
     def __init__(self, enabled=True, options=None):
         if options is not None:
             self.profiler_options = options
diff --git a/python/paddle/vision/transforms/transforms.py b/python/paddle/vision/transforms/transforms.py
index 78413da4f208554a446ba4b5d2e72fadeef549f2..f5cbd90ffc7b0d8a5023f42122f4bd463f552c7e 100644
--- a/python/paddle/vision/transforms/transforms.py
+++ b/python/paddle/vision/transforms/transforms.py
@@ -84,7 +84,7 @@ def _check_input(
     return value
 
 
-class Compose(object):
+class Compose:
     """
     Composes several transforms together use for composing list of transforms
     together for a dataset transform.
@@ -137,7 +137,7 @@ class Compose(object):
         return format_string
 
 
-class BaseTransform(object):
+class BaseTransform:
     """
     Base class of all transforms used in computer vision.
 
diff --git a/tools/CrossStackProfiler/CspChromeTraceFormatter.py b/tools/CrossStackProfiler/CspChromeTraceFormatter.py
index 1fa8efe9880508dbbedca02541337d7f795a8afa..fb24e0634e546c587fb582637908eafce5970c49 100755
--- a/tools/CrossStackProfiler/CspChromeTraceFormatter.py
+++ b/tools/CrossStackProfiler/CspChromeTraceFormatter.py
@@ -15,7 +15,7 @@
 import json
 
 
-class ChromeTraceFormatter(object):
+class ChromeTraceFormatter:
     def __init__(self):
         self._events = []
         self._metadata = []
diff --git a/tools/CrossStackProfiler/CspFileReader.py b/tools/CrossStackProfiler/CspFileReader.py
index 11dd052283c1564f0cb116787f268bd69f12fce8..55a1722be66b7cc7683cbda06d783cc2e1787002 100755
--- a/tools/CrossStackProfiler/CspFileReader.py
+++ b/tools/CrossStackProfiler/CspFileReader.py
@@ -73,7 +73,7 @@ FILEORGANIZEFORM = [
 ]
 
 
-class FileReader(object):
+class FileReader:
     def __init__(self, logger, args):
         self._logger = logger
         self._args = args
diff --git a/tools/CrossStackProfiler/CspReporter.py b/tools/CrossStackProfiler/CspReporter.py
index 052ffd6fca19f4b54a5c9c797703f2045275f6a6..999ba4fb3e23368151732e7d249c76bbac223f59 100755
--- a/tools/CrossStackProfiler/CspReporter.py
+++ b/tools/CrossStackProfiler/CspReporter.py
@@ -64,7 +64,7 @@ def get_argparse():
     return parser.parse_args()
 
 
-class CspReporter(object):
+class CspReporter:
     def __init__(self, args):
         self._args = args
         print(self._args)
diff --git a/tools/check_ut.py b/tools/check_ut.py
index e08c358531f7fd17a27de70fae4260d0c82b92fe..6f1a8ab02ba61e53e2c671bd42294842ddb8db71 100644
--- a/tools/check_ut.py
+++ b/tools/check_ut.py
@@ -19,7 +19,7 @@ import os.path
 from github import Github
 
 
-class PRChecker(object):
+class PRChecker:
     """PR Checker."""
 
     def __init__(self):
diff --git a/tools/codestyle/docstring_checker.py b/tools/codestyle/docstring_checker.py
index 0d163c20bbfa15474e568b2251a216370659da58..8deeff77348f476e7e5d29216ae0de298beba212 100644
--- a/tools/codestyle/docstring_checker.py
+++ b/tools/codestyle/docstring_checker.py
@@ -27,7 +27,7 @@ def register(linter):
     linter.register_checker(DocstringChecker(linter))
 
 
-class Docstring(object):
+class Docstring:
     """Docstring class holds the parsed doc string elements."""
 
     def __init__(self):
diff --git a/tools/get_pr_ut.py b/tools/get_pr_ut.py
index d3a1c41682bd9b2a54fdd36682f9e50765e07c5b..02aaf13a17d31f1d0a7e6dfdfa4c90b53c35e8cb 100644
--- a/tools/get_pr_ut.py
+++ b/tools/get_pr_ut.py
@@ -30,7 +30,7 @@ PADDLE_ROOT = PADDLE_ROOT.replace('//', '/')
 ssl._create_default_https_context = ssl._create_unverified_context
 
 
-class PRChecker(object):
+class PRChecker:
     """PR Checker."""
 
     def __init__(self):
diff --git a/tools/timeline.py b/tools/timeline.py
index 5323ea0f907ec603a310538306229fc7db1b1606..a8c6699b270d049f4844cd98a0da60377d6e814e 100644
--- a/tools/timeline.py
+++ b/tools/timeline.py
@@ -31,7 +31,7 @@ parser.add_argument(
 args = parser.parse_args()
 
 
-class _ChromeTraceFormatter(object):
+class _ChromeTraceFormatter:
     def __init__(self):
         self._events = []
         self._metadata = []
@@ -126,7 +126,7 @@ class _ChromeTraceFormatter(object):
             return json.dumps(trace, separators=(',', ':'))
 
 
-class Timeline(object):
+class Timeline:
     def __init__(self, profile_dict):
         self._profile_dict = profile_dict
         self._pid = 0