diff --git a/core/engine/cluster/cloud/__init__.py b/core/engine/cluster/cloud/__init__.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..abf198b97e6e818e1fbe59006f98492640bcee54 100644
--- a/core/engine/cluster/cloud/__init__.py
+++ b/core/engine/cluster/cloud/__init__.py
@@ -0,0 +1,13 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/core/modules/coding/layers.py b/core/modules/coding/layers.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..abf198b97e6e818e1fbe59006f98492640bcee54 100755
--- a/core/modules/coding/layers.py
+++ b/core/modules/coding/layers.py
@@ -0,0 +1,13 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/core/trainers/__init__.py b/core/trainers/__init__.py
index cd9c9db5e6b93fd6171bca0a5b0f97f69306aedc..f14704cad8f3859746f95353ba68753f857ff78d 100755
--- a/core/trainers/__init__.py
+++ b/core/trainers/__init__.py
@@ -11,7 +11,6 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
 """
 trainer implement.
 
@@ -22,5 +21,3 @@ Trainer
                                ↘  (for online learning training) OnlineLearningTrainer
 
 """
-
-
diff --git a/core/trainers/ctr_coding_trainer.py b/core/trainers/ctr_coding_trainer.py
index 3bfec28cfd149bdbe47fdc202107c7ed7af58fdd..7dc51f340147aec933ce8bffd0be080b7be984c6 100755
--- a/core/trainers/ctr_coding_trainer.py
+++ b/core/trainers/ctr_coding_trainer.py
@@ -59,8 +59,10 @@ class CtrTrainer(Trainer):
         reader_class = envs.get_global_env("class", None, namespace)
         abs_dir = os.path.dirname(os.path.abspath(__file__))
         reader = os.path.join(abs_dir, '../utils', 'dataset_instance.py')
-        pipe_cmd = "python {} {} {} {}".format(reader, reader_class, "TRAIN", self._config_yaml)
-        train_data_path = envs.get_global_env("train_data_path", None, namespace)
+        pipe_cmd = "python {} {} {} {}".format(reader, reader_class, "TRAIN",
+                                               self._config_yaml)
+        train_data_path = envs.get_global_env("train_data_path", None,
+                                              namespace)
 
         dataset = fluid.DatasetFactory().create_dataset()
         dataset.set_use_var(inputs)
@@ -87,7 +89,8 @@ class CtrTrainer(Trainer):
         self.model.train_net()
         optimizer = self.model.optimizer()
 
-        optimizer = fleet.distributed_optimizer(optimizer, strategy={"use_cvm": False})
+        optimizer = fleet.distributed_optimizer(
+            optimizer, strategy={"use_cvm": False})
         optimizer.minimize(self.model.get_avg_cost())
 
         if fleet.is_server():
@@ -118,16 +121,18 @@ class CtrTrainer(Trainer):
         gs = shuf * 0
         fleet._role_maker._node_type_comm.Allreduce(shuf, gs)
 
-        print("trainer id: {}, trainers: {}, gs: {}".format(fleet.worker_index(), fleet.worker_num(), gs))
+        print("trainer id: {}, trainers: {}, gs: {}".format(fleet.worker_index(
+        ), fleet.worker_num(), gs))
 
         epochs = envs.get_global_env("train.epochs")
 
         for i in range(epochs):
-            self._exe.train_from_dataset(program=fluid.default_main_program(),
-                                         dataset=dataset,
-                                         fetch_list=self.fetch_vars,
-                                         fetch_info=self.fetch_alias,
-                                         print_period=self.fetch_period)
+            self._exe.train_from_dataset(
+                program=fluid.default_main_program(),
+                dataset=dataset,
+                fetch_list=self.fetch_vars,
+                fetch_info=self.fetch_alias,
+                print_period=self.fetch_period)
 
         context['status'] = 'terminal_pass'
         fleet.stop_worker()
diff --git a/core/trainers/ctr_modul_trainer.py b/core/trainers/ctr_modul_trainer.py
index 7b3bd7874359059c3b03289cc10da7d7756ac35b..af8f3f3a2c3fb59fc6db60e3e4cd050ca3d8ad8a 100755
--- a/core/trainers/ctr_modul_trainer.py
+++ b/core/trainers/ctr_modul_trainer.py
@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-
 import datetime
 import json
 import sys
@@ -23,7 +22,6 @@ import paddle.fluid as fluid
 from paddle.fluid.incubate.fleet.parameter_server.pslib import fleet
 from paddle.fluid.incubate.fleet.base.role_maker import GeneralRoleMaker
 
-
 from paddlerec.core.utils import fs as fs
 from paddlerec.core.utils import util as util
 from paddlerec.core.metrics.auc_metrics import AUCMetric
@@ -80,20 +78,31 @@ class CtrTrainer(Trainer):
         """R
         """
         Trainer.__init__(self, config)
-        config['output_path'] = util.get_absolute_path(
-            config['output_path'], config['io']['afs'])
+        config['output_path'] = util.get_absolute_path(config['output_path'],
+                                                       config['io']['afs'])
 
         self.global_config = config
         self._metrics = {}
 
         self._path_generator = util.PathGenerator({
-            'templates': [
-                {'name': 'xbox_base_done', 'template': config['output_path'] + '/xbox_base_done.txt'},
-                {'name': 'xbox_delta_done', 'template': config['output_path'] + '/xbox_patch_done.txt'},
-                {'name': 'xbox_base', 'template': config['output_path'] + '/xbox/{day}/base/'},
-                {'name': 'xbox_delta', 'template': config['output_path'] + '/xbox/{day}/delta-{pass_id}/'},
-                {'name': 'batch_model', 'template': config['output_path'] + '/batch_model/{day}/{pass_id}/'}
-            ]
+            'templates': [{
+                'name': 'xbox_base_done',
+                'template': config['output_path'] + '/xbox_base_done.txt'
+            }, {
+                'name': 'xbox_delta_done',
+                'template': config['output_path'] + '/xbox_patch_done.txt'
+            }, {
+                'name': 'xbox_base',
+                'template': config['output_path'] + '/xbox/{day}/base/'
+            }, {
+                'name': 'xbox_delta',
+                'template':
+                config['output_path'] + '/xbox/{day}/delta-{pass_id}/'
+            }, {
+                'name': 'batch_model',
+                'template':
+                config['output_path'] + '/batch_model/{day}/{pass_id}/'
+            }]
         })
         if 'path_generator' in config:
             self._path_generator.add_path_template(config['path_generator'])
@@ -111,9 +120,11 @@ class CtrTrainer(Trainer):
         if self.global_config.get('process_mode', 'mpi') == 'brilliant_cpu':
             afs_config = self.global_config['io']['afs']
             role_maker = GeneralRoleMaker(
-                hdfs_name=afs_config['fs_name'], hdfs_ugi=afs_config['fs_ugi'],
+                hdfs_name=afs_config['fs_name'],
+                hdfs_ugi=afs_config['fs_ugi'],
                 path=self.global_config['output_path'] + "/gloo",
-                init_timeout_seconds=1200, run_timeout_seconds=1200)
+                init_timeout_seconds=1200,
+                run_timeout_seconds=1200)
         fleet.init(role_maker)
         data_var_list = []
         data_var_name_dict = {}
@@ -125,7 +136,8 @@ class CtrTrainer(Trainer):
             scope = fluid.Scope()
             self._exector_context[executor['name']] = {}
             self._exector_context[executor['name']]['scope'] = scope
-            self._exector_context[executor['name']]['model'] = model_basic.create(executor)
+            self._exector_context[executor['name']][
+                'model'] = model_basic.create(executor)
             model = self._exector_context[executor['name']]['model']
             self._metrics.update(model.get_metrics())
             runnnable_scope.append(scope)
@@ -146,9 +158,12 @@ class CtrTrainer(Trainer):
             model = self._exector_context[executor['name']]['model']
             program = model._build_param['model']['train_program']
             if not executor['is_update_sparse']:
-                program._fleet_opt["program_configs"][str(id(model.get_avg_cost().block.program))]["push_sparse"] = []
+                program._fleet_opt["program_configs"][str(
+                    id(model.get_avg_cost().block.program))][
+                        "push_sparse"] = []
             if 'train_thread_num' not in executor:
-                executor['train_thread_num'] = self.global_config['train_thread_num']
+                executor['train_thread_num'] = self.global_config[
+                    'train_thread_num']
             with fluid.scope_guard(scope):
                 self._exe.run(model._build_param['model']['startup_program'])
             model.dump_model_program('./')
@@ -162,7 +177,8 @@ class CtrTrainer(Trainer):
             dataset_item['data_vars'] = data_var_list
             dataset_item.update(self.global_config['io']['afs'])
             dataset_item["batch_size"] = self.global_config['batch_size']
-            self._dataset[dataset_item['name']] = dataset.FluidTimeSplitDataset(dataset_item)
+            self._dataset[dataset_item[
+                'name']] = dataset.FluidTimeSplitDataset(dataset_item)
         # if config.need_reqi_changeslot and config.reqi_dnn_plugin_day >= last_day and config.reqi_dnn_plugin_pass >= last_pass:
         #    util.reqi_changeslot(config.hdfs_dnn_plugin_path, join_save_params, common_save_params, update_save_params, scope2, scope3)
         fleet.init_worker()
@@ -190,23 +206,30 @@ class CtrTrainer(Trainer):
             metric_param = {'label': metric, 'metric_dict': metrics[metric]}
             metric_calculator.calculate(scope, metric_param)
             metric_result = metric_calculator.get_result_to_string()
-            self.print_log(metric_result, {'master': True, 'stdout': stdout_str})
+            self.print_log(metric_result,
+                           {'master': True,
+                            'stdout': stdout_str})
             monitor_data += metric_result
             metric_calculator.clear(scope, metric_param)
 
     def save_model(self, day, pass_index, base_key):
         """R
         """
-        cost_printer = util.CostPrinter(util.print_cost,
-                                        {'master': True, 'log_format': 'save model cost %s sec'})
-        model_path = self._path_generator.generate_path('batch_model', {'day': day, 'pass_id': pass_index})
+        cost_printer = util.CostPrinter(util.print_cost, {
+            'master': True,
+            'log_format': 'save model cost %s sec'
+        })
+        model_path = self._path_generator.generate_path(
+            'batch_model', {'day': day,
+                            'pass_id': pass_index})
         save_mode = 0  # just save all
         if pass_index < 1:  # batch_model
             save_mode = 3  # unseen_day++, save all
         util.rank0_print("going to save_model %s" % model_path)
         fleet.save_persistables(None, model_path, mode=save_mode)
         if fleet._role_maker.is_first_worker():
-            self._train_pass.save_train_progress(day, pass_index, base_key, model_path, is_checkpoint=True)
+            self._train_pass.save_train_progress(
+                day, pass_index, base_key, model_path, is_checkpoint=True)
         cost_printer.done()
         return model_path
 
@@ -225,46 +248,58 @@ class CtrTrainer(Trainer):
         if pass_index < 1:
             save_mode = 2
             xbox_patch_id = xbox_base_key
-            model_path = self._path_generator.generate_path('xbox_base', {'day': day})
-            xbox_model_donefile = self._path_generator.generate_path('xbox_base_done', {'day': day})
+            model_path = self._path_generator.generate_path('xbox_base',
+                                                            {'day': day})
+            xbox_model_donefile = self._path_generator.generate_path(
+                'xbox_base_done', {'day': day})
         else:
             save_mode = 1
-            model_path = self._path_generator.generate_path('xbox_delta', {'day': day, 'pass_id': pass_index})
-            xbox_model_donefile = self._path_generator.generate_path('xbox_delta_done', {'day': day})
-        total_save_num = fleet.save_persistables(None, model_path, mode=save_mode)
+            model_path = self._path_generator.generate_path(
+                'xbox_delta', {'day': day,
+                               'pass_id': pass_index})
+            xbox_model_donefile = self._path_generator.generate_path(
+                'xbox_delta_done', {'day': day})
+        total_save_num = fleet.save_persistables(
+            None, model_path, mode=save_mode)
         cost_printer.done()
 
-        cost_printer = util.CostPrinter(util.print_cost, {'master': True,
-                                                          'log_format': 'save cache model cost %s sec',
-                                                          'stdout': stdout_str})
+        cost_printer = util.CostPrinter(util.print_cost, {
+            'master': True,
+            'log_format': 'save cache model cost %s sec',
+            'stdout': stdout_str
+        })
         model_file_handler = fs.FileHandler(self.global_config['io']['afs'])
         if self.global_config['save_cache_model']:
-            cache_save_num = fleet.save_cache_model(None, model_path, mode=save_mode)
+            cache_save_num = fleet.save_cache_model(
+                None, model_path, mode=save_mode)
             model_file_handler.write(
                 "file_prefix:part\npart_num:16\nkey_num:%d\n" % cache_save_num,
                 model_path + '/000_cache/sparse_cache.meta', 'w')
         cost_printer.done()
-        util.rank0_print("save xbox cache model done, key_num=%s" % cache_save_num)
+        util.rank0_print("save xbox cache model done, key_num=%s" %
+                         cache_save_num)
 
-        save_env_param = {
-            'executor': self._exe,
-            'save_combine': True
-        }
-        cost_printer = util.CostPrinter(util.print_cost, {'master': True,
-                                                          'log_format': 'save dense model cost %s sec',
-                                                          'stdout': stdout_str})
+        save_env_param = {'executor': self._exe, 'save_combine': True}
+        cost_printer = util.CostPrinter(util.print_cost, {
+            'master': True,
+            'log_format': 'save dense model cost %s sec',
+            'stdout': stdout_str
+        })
         if fleet._role_maker.is_first_worker():
             for executor in self.global_config['executor']:
                 if 'layer_for_inference' not in executor:
                     continue
                 executor_name = executor['name']
                 model = self._exector_context[executor_name]['model']
-                save_env_param['inference_list'] = executor['layer_for_inference']
-                save_env_param['scope'] = self._exector_context[executor_name]['scope']
+                save_env_param['inference_list'] = executor[
+                    'layer_for_inference']
+                save_env_param['scope'] = self._exector_context[executor_name][
+                    'scope']
                 model.dump_inference_param(save_env_param)
                 for dnn_layer in executor['layer_for_inference']:
                     model_file_handler.cp(dnn_layer['save_file_name'],
-                                          model_path + '/dnn_plugin/' + dnn_layer['save_file_name'])
+                                          model_path + '/dnn_plugin/' +
+                                          dnn_layer['save_file_name'])
         fleet._role_maker._barrier_worker()
         cost_printer.done()
 
@@ -282,9 +317,15 @@ class CtrTrainer(Trainer):
             "job_name": util.get_env_value("JOB_NAME")
         }
         if fleet._role_maker.is_first_worker():
-            model_file_handler.write(json.dumps(xbox_done_info) + "\n", xbox_model_donefile, 'a')
+            model_file_handler.write(
+                json.dumps(xbox_done_info) + "\n", xbox_model_donefile, 'a')
             if pass_index > 0:
-                self._train_pass.save_train_progress(day, pass_index, xbox_base_key, model_path, is_checkpoint=False)
+                self._train_pass.save_train_progress(
+                    day,
+                    pass_index,
+                    xbox_base_key,
+                    model_path,
+                    is_checkpoint=False)
         fleet._role_maker._barrier_worker()
         return stdout_str
 
@@ -301,21 +342,28 @@ class CtrTrainer(Trainer):
             util.rank0_print("Begin " + executor_name + " pass")
             begin = time.time()
             program = model._build_param['model']['train_program']
-            self._exe.train_from_dataset(program, dataset, scope,
-                                         thread=executor_config['train_thread_num'], debug=self.global_config['debug'])
+            self._exe.train_from_dataset(
+                program,
+                dataset,
+                scope,
+                thread=executor_config['train_thread_num'],
+                debug=self.global_config['debug'])
             end = time.time()
             local_cost = (end - begin) / 60.0
             avg_cost = worker_numric_avg(local_cost)
             min_cost = worker_numric_min(local_cost)
             max_cost = worker_numric_max(local_cost)
-            util.rank0_print("avg train time %s mins, min %s mins, max %s mins" % (avg_cost, min_cost, max_cost))
+            util.rank0_print("avg train time %s mins, min %s mins, max %s mins"
+                             % (avg_cost, min_cost, max_cost))
             self._exector_context[executor_name]['cost'] = max_cost
 
             monitor_data = ""
             self.print_global_metrics(scope, model, monitor_data, stdout_str)
             util.rank0_print("End " + executor_name + " pass")
-            if self._train_pass.need_dump_inference(pass_id) and executor_config['dump_inference_model']:
-                stdout_str += self.save_xbox_model(day, pass_id, xbox_base_key, monitor_data)
+            if self._train_pass.need_dump_inference(
+                    pass_id) and executor_config['dump_inference_model']:
+                stdout_str += self.save_xbox_model(day, pass_id, xbox_base_key,
+                                                   monitor_data)
         fleet._role_maker._barrier_worker()
 
     def startup(self, context):
@@ -328,10 +376,14 @@ class CtrTrainer(Trainer):
         stdout_str = ""
         self._train_pass = util.TimeTrainPass(self.global_config)
         if not self.global_config['cold_start']:
-            cost_printer = util.CostPrinter(util.print_cost,
-                                            {'master': True, 'log_format': 'load model cost %s sec',
-                                             'stdout': stdout_str})
-            self.print_log("going to load model %s" % self._train_pass._checkpoint_model_path, {'master': True})
+            cost_printer = util.CostPrinter(util.print_cost, {
+                'master': True,
+                'log_format': 'load model cost %s sec',
+                'stdout': stdout_str
+            })
+            self.print_log("going to load model %s" %
+                           self._train_pass._checkpoint_model_path,
+                           {'master': True})
             # if config.need_reqi_changeslot and config.reqi_dnn_plugin_day >= self._train_pass.date()
             #    and config.reqi_dnn_plugin_pass >= self._pass_id:
             #    fleet.load_one_table(0, self._train_pass._checkpoint_model_path)
@@ -340,9 +392,12 @@ class CtrTrainer(Trainer):
             cost_printer.done()
         if self.global_config['save_first_base']:
             self.print_log("save_first_base=True", {'master': True})
-            self.print_log("going to save xbox base model", {'master': True, 'stdout': stdout_str})
+            self.print_log("going to save xbox base model",
+                           {'master': True,
+                            'stdout': stdout_str})
             self._train_pass._base_key = int(time.time())
-            stdout_str += self.save_xbox_model(self._train_pass.date(), 0, self._train_pass._base_key, "")
+            stdout_str += self.save_xbox_model(self._train_pass.date(), 0,
+                                               self._train_pass._base_key, "")
         context['status'] = 'begin_day'
 
     def begin_day(self, context):
@@ -353,7 +408,9 @@ class CtrTrainer(Trainer):
             context['is_exit'] = True
         day = self._train_pass.date()
         pass_id = self._train_pass._pass_id
-        self.print_log("======== BEGIN DAY:%s ========" % day, {'master': True, 'stdout': stdout_str})
+        self.print_log("======== BEGIN DAY:%s ========" % day,
+                       {'master': True,
+                        'stdout': stdout_str})
         if pass_id == self._train_pass.max_pass_num_day():
             context['status'] = 'end_day'
         else:
@@ -368,8 +425,10 @@ class CtrTrainer(Trainer):
         context['status'] = 'begin_day'
 
         util.rank0_print("shrink table")
-        cost_printer = util.CostPrinter(util.print_cost,
-                                        {'master': True, 'log_format': 'shrink table done, cost %s sec'})
+        cost_printer = util.CostPrinter(util.print_cost, {
+            'master': True,
+            'log_format': 'shrink table done, cost %s sec'
+        })
         fleet.shrink_sparse_table()
         for executor in self._exector_context:
             self._exector_context[executor]['model'].shrink({
@@ -394,7 +453,9 @@ class CtrTrainer(Trainer):
         pass_id = self._train_pass._pass_id
         base_key = self._train_pass._base_key
         pass_time = self._train_pass._current_train_time.strftime("%Y%m%d%H%M")
-        self.print_log("    ==== begin delta:%s ========" % pass_id, {'master': True, 'stdout': stdout_str})
+        self.print_log("    ==== begin delta:%s ========" % pass_id,
+                       {'master': True,
+                        'stdout': stdout_str})
         train_begin_time = time.time()
 
         cost_printer = util.CostPrinter(util.print_cost, \
@@ -403,35 +464,46 @@ class CtrTrainer(Trainer):
         current_dataset = {}
         for name in self._dataset:
             current_dataset[name] = self._dataset[name].load_dataset({
-                'node_num': fleet.worker_num(), 'node_idx': fleet.worker_index(),
-                'begin_time': pass_time, 'time_window_min': self._train_pass._interval_per_pass
+                'node_num': fleet.worker_num(),
+                'node_idx': fleet.worker_index(),
+                'begin_time': pass_time,
+                'time_window_min': self._train_pass._interval_per_pass
             })
         fleet._role_maker._barrier_worker()
         cost_printer.done()
 
         util.rank0_print("going to global shuffle")
         cost_printer = util.CostPrinter(util.print_cost, {
-            'master': True, 'stdout': stdout_str,
-            'log_format': 'global shuffle done, cost %s sec'})
+            'master': True,
+            'stdout': stdout_str,
+            'log_format': 'global shuffle done, cost %s sec'
+        })
         for name in current_dataset:
-            current_dataset[name].global_shuffle(fleet, self.global_config['dataset']['shuffle_thread'])
+            current_dataset[name].global_shuffle(
+                fleet, self.global_config['dataset']['shuffle_thread'])
         cost_printer.done()
         # str(dataset.get_shuffle_data_size(fleet))
         fleet._role_maker._barrier_worker()
 
         if self.global_config['prefetch_data']:
-            next_pass_time = (self._train_pass._current_train_time +
-                              datetime.timedelta(minutes=self._train_pass._interval_per_pass)).strftime("%Y%m%d%H%M")
+            next_pass_time = (
+                self._train_pass._current_train_time + datetime.timedelta(
+                    minutes=self._train_pass._interval_per_pass)
+            ).strftime("%Y%m%d%H%M")
             for name in self._dataset:
                 self._dataset[name].preload_dataset({
-                    'node_num': fleet.worker_num(), 'node_idx': fleet.worker_index(),
-                    'begin_time': next_pass_time, 'time_window_min': self._train_pass._interval_per_pass
+                    'node_num': fleet.worker_num(),
+                    'node_idx': fleet.worker_index(),
+                    'begin_time': next_pass_time,
+                    'time_window_min': self._train_pass._interval_per_pass
                 })
 
         fleet._role_maker._barrier_worker()
         pure_train_begin = time.time()
         for executor in self.global_config['executor']:
-            self.run_executor(executor, current_dataset[executor['dataset_name']], stdout_str)
+            self.run_executor(executor,
+                              current_dataset[executor['dataset_name']],
+                              stdout_str)
         cost_printer = util.CostPrinter(util.print_cost, \
                                         {'master': True, 'log_format': 'release_memory cost %s sec'})
         for name in current_dataset:
@@ -444,9 +516,11 @@ class CtrTrainer(Trainer):
         train_end_time = time.time()
         train_cost = train_end_time - train_begin_time
         other_cost = train_cost - pure_train_cost
-        log_str = "finished train day %s pass %s time cost:%s sec job time cost:" % (day, pass_id, train_cost)
+        log_str = "finished train day %s pass %s time cost:%s sec job time cost:" % (
+            day, pass_id, train_cost)
         for executor in self._exector_context:
-            log_str += '[' + executor + ':' + str(self._exector_context[executor]['cost']) + ']'
+            log_str += '[' + executor + ':' + str(self._exector_context[
+                executor]['cost']) + ']'
         log_str += '[other_cost:' + str(other_cost) + ']'
         util.rank0_print(log_str)
         stdout_str += util.now_time_str() + log_str
diff --git a/doc/benchmark.md b/doc/benchmark.md
index b16e26c71888d590f00f13782449acf840c4b6ee..2aaea25d6941043d24fcee31c7117a3b34c4f525 100644
--- a/doc/benchmark.md
+++ b/doc/benchmark.md
@@ -1,2 +1,2 @@
 # PaddleRec Benchmark
-> 占位
\ No newline at end of file
+> 占位
diff --git a/doc/contribute.md b/doc/contribute.md
index a9bd1910021e78573f1d9fd99f66404b77927737..26770d8ac0b64e9835f7398768d0f81de9383132 100644
--- a/doc/contribute.md
+++ b/doc/contribute.md
@@ -1,2 +1,2 @@
 # PaddleRec 贡献代码
-> 占位
\ No newline at end of file
+> 占位
diff --git a/doc/design.md b/doc/design.md
index 2975d77f14e461547921f74b9ced5cf73703e2e7..a442bd16a25301178538f482cd537a4ca23bc395 100644
--- a/doc/design.md
+++ b/doc/design.md
@@ -279,4 +279,4 @@ class Metric(object):
         pass
 ```
 
-全局指标的计算及输出，需要分别继承并实现以上四个成员函数。具体实现的例子，可以参考[auc_metric.py](../core/metrics/auc_metrics.py)
\ No newline at end of file
+全局指标的计算及输出，需要分别继承并实现以上四个成员函数。具体实现的例子，可以参考[auc_metric.py](../core/metrics/auc_metrics.py)
diff --git a/doc/distributed_train.md b/doc/distributed_train.md
index 425f141ab76e173a9484dff90cd5cfb55acaf853..339c5a83ffd26f9416a67a02390a11ba4c87c29d 100644
--- a/doc/distributed_train.md
+++ b/doc/distributed_train.md
@@ -7,5 +7,3 @@
 
 ### K8S集群运行分布式
 > 占位
-
-
diff --git a/doc/faq.md b/doc/faq.md
index f7ca7cc4a7c366a2a828496eae3f12d1dea17b7e..60790140877b6b11add29552e02c0a435da75f87 100644
--- a/doc/faq.md
+++ b/doc/faq.md
@@ -1,2 +1,2 @@
 # 常见问题FAQ
-> 占位
\ No newline at end of file
+> 占位
diff --git a/doc/local_train.md b/doc/local_train.md
index 4a43fa5520ca745badc4d2a49710763eac6e7a0a..e65255ebf7e14933f52f9977b2ecec48dabbb76e 100644
--- a/doc/local_train.md
+++ b/doc/local_train.md
@@ -1,2 +1,2 @@
 # PaddleRec 单机训练
-> 占位
\ No newline at end of file
+> 占位
diff --git a/doc/model_list.md b/doc/model_list.md
index 9e68d9f6d2f8e9361cc13b9e76f28426062943bc..b46687a60475fbd309f01050194510b21b060f17 100644
--- a/doc/model_list.md
+++ b/doc/model_list.md
@@ -12,4 +12,3 @@
 |  多任务  |        [ESMM]()        |      ✓      |      x      |       ✓       |     x      |       ✓       |      ✓       |
 |   匹配   |        [DSSM]()        |      ✓      |      x      |       ✓       |     x      |       ✓       |      ✓       |
 |   匹配   |  [Multiview-Simnet]()  |      ✓      |      x      |       ✓       |     x      |       ✓       |      ✓       |
-
diff --git a/doc/optimization_model.md b/doc/optimization_model.md
index b516f8958053b1b2bd6982b71c699e8baf69f8d9..e63f45b62b50db55f1c6c0d48c7ca23b016b74d3 100644
--- a/doc/optimization_model.md
+++ b/doc/optimization_model.md
@@ -1,2 +1,2 @@
 # PaddleRec 模型调参
-> 占位
\ No newline at end of file
+> 占位
diff --git a/doc/predict.md b/doc/predict.md
index a33eda43ec6aed8ebe628f0540327b707055970d..07160e1f0e7563276c33e514d006dd3747492f90 100644
--- a/doc/predict.md
+++ b/doc/predict.md
@@ -1 +1 @@
-# PaddleRec 离线预测
\ No newline at end of file
+# PaddleRec 离线预测
diff --git a/doc/ps_background.md b/doc/ps_background.md
index 984e1b00c96242843cceaf68cf15bb8deb52c391..e5f2e320940763986351fefd21a5e1f1363b6104 100644
--- a/doc/ps_background.md
+++ b/doc/ps_background.md
@@ -5,4 +5,3 @@
 
 
 ## [参数服务器训练](https://www.paddlepaddle.org.cn/tutorials/projectdetail/464839)
-
diff --git a/models/contentunderstanding/__init__.py b/models/contentunderstanding/__init__.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..abf198b97e6e818e1fbe59006f98492640bcee54 100755
--- a/models/contentunderstanding/__init__.py
+++ b/models/contentunderstanding/__init__.py
@@ -0,0 +1,13 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/models/contentunderstanding/classification/config.yaml b/models/contentunderstanding/classification/config.yaml
index d1748137f0c4d994b3a566debf43dbdc2c3d66dc..ef55cd18e8fd45829acd2f479c661f27decfda71 100644
--- a/models/contentunderstanding/classification/config.yaml
+++ b/models/contentunderstanding/classification/config.yaml
@@ -37,4 +37,3 @@ train:
       dirname: "inference"
       epoch_interval: 100
       save_last: True
-
diff --git a/models/contentunderstanding/classification/model.py b/models/contentunderstanding/classification/model.py
index 9e853aa01d4a0b6bd5c7a20d8e13164bd9905ad0..23c51d44d7d839d9db30f8129c3e42449a6a80d4 100644
--- a/models/contentunderstanding/classification/model.py
+++ b/models/contentunderstanding/classification/model.py
@@ -31,7 +31,8 @@ class Model(ModelBase):
     def train_net(self):
         """ network definition """
 
-        data = fluid.data(name="input", shape=[None, self.max_len], dtype='int64')
+        data = fluid.data(
+            name="input", shape=[None, self.max_len], dtype='int64')
         label = fluid.data(name="label", shape=[None, 1], dtype='int64')
         seq_len = fluid.data(name="seq_len", shape=[None], dtype='int64')
 
@@ -51,7 +52,9 @@ class Model(ModelBase):
         # full connect layer
         fc_1 = fluid.layers.fc(input=[conv], size=self.hid_dim)
         # softmax layer
-        prediction = fluid.layers.fc(input=[fc_1], size=self.class_dim, act="softmax")
+        prediction = fluid.layers.fc(input=[fc_1],
+                                     size=self.class_dim,
+                                     act="softmax")
         cost = fluid.layers.cross_entropy(input=prediction, label=label)
         avg_cost = fluid.layers.mean(x=cost)
         acc = fluid.layers.accuracy(input=prediction, label=label)
diff --git a/models/contentunderstanding/classification/reader.py b/models/contentunderstanding/classification/reader.py
index 136a5668856c0fb558a016a3bc3a0b8a56651d3b..1c8e86cdb49f1cc89c9c4f413cbd7b117b55aa55 100644
--- a/models/contentunderstanding/classification/reader.py
+++ b/models/contentunderstanding/classification/reader.py
@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-
 import sys
 
 from paddlerec.core.reader import Reader
@@ -38,7 +37,8 @@ class TrainReader(Reader):
             data = [int(i) for i in data]
             label = [int(i) for i in label]
             seq_len = [int(i) for i in seq_len]
-            print >> sys.stderr, str([('data', data), ('label', label), ('seq_len', seq_len)])
+            print >> sys.stderr, str(
+                [('data', data), ('label', label), ('seq_len', seq_len)])
             yield [('data', data), ('label', label), ('seq_len', seq_len)]
 
         return data_iter
diff --git a/models/contentunderstanding/readme.md b/models/contentunderstanding/readme.md
index 07ec96f2414881998617f048c1e82a3f0d9cda75..deefbd2eb02f08d7fac810eb40ae78ff1a173baf 100644
--- a/models/contentunderstanding/readme.md
+++ b/models/contentunderstanding/readme.md
@@ -87,4 +87,3 @@ python -m paddlerec.run -m paddlerec.models.contentunderstanding.classification
 | :------------------: | :--------------------: | :---------: |:---------: | :---------: |:---------: |
 |       ag news dataset        |       TagSpace       |       --        |       --          |       --          |       --          |
 |       --        |       Classification       |       --        |       --          |       --          |       --          |
-
diff --git a/models/contentunderstanding/tagspace/config.yaml b/models/contentunderstanding/tagspace/config.yaml
index 70333fcbf7edf4b6b5f54145e29cb122ed3ae9c6..19fbf277d66445c44287856512cb0b13777dc251 100644
--- a/models/contentunderstanding/tagspace/config.yaml
+++ b/models/contentunderstanding/tagspace/config.yaml
@@ -47,4 +47,3 @@ train:
       dirname: "inference"
       epoch_interval: 100
       save_last: True
-
diff --git a/models/match/__init__.py b/models/match/__init__.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..abf198b97e6e818e1fbe59006f98492640bcee54 100755
--- a/models/match/__init__.py
+++ b/models/match/__init__.py
@@ -0,0 +1,13 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/models/match/readme.md b/models/match/readme.md
index 6bccc109ff14582e816dee64b72b786a1e90f49e..d9f91b257d81ffde820a04cad49b56edbd903f6a 100755
--- a/models/match/readme.md
+++ b/models/match/readme.md
@@ -37,4 +37,3 @@
 python -m paddlerec.run -m paddlerec.models.match.dssm # dssm
 python -m paddlerec.run -m paddlerec.models.match.multiview-simnet # multiview-simnet
 ```
-
diff --git a/models/multitask/__init__.py b/models/multitask/__init__.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..abf198b97e6e818e1fbe59006f98492640bcee54 100755
--- a/models/multitask/__init__.py
+++ b/models/multitask/__init__.py
@@ -0,0 +1,13 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/models/multitask/readme.md b/models/multitask/readme.md
index d234f42f146e18bf254e518db0e78acc1e1d3e10..10e0641060f74b67b4987d14a1c4aad27a25b103 100755
--- a/models/multitask/readme.md
+++ b/models/multitask/readme.md
@@ -56,4 +56,3 @@ python -m paddlerec.run -m paddlerec.models.multitask.esmm # esmm
 |       Census-income Data     |       Share-Bottom       |       --        |     0.93120/0.99256         |
 |       Census-income Data        |       MMoE       |       --        |       0.94465/0.99324         |
 |          Ali-CCP     |    ESMM       |       --        |      0.97181/0.49967          |
-
diff --git a/models/rank/__init__.py b/models/rank/__init__.py
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..abf198b97e6e818e1fbe59006f98492640bcee54 100755
--- a/models/rank/__init__.py
+++ b/models/rank/__init__.py
@@ -0,0 +1,13 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/models/rank/dcn/data/download.py b/models/rank/dcn/data/download.py
index d9bcc6df296068cfd5cd9fc1c91165f11b580d04..4203a3868a577757930ae848736c34bb4da376c7 100755
--- a/models/rank/dcn/data/download.py
+++ b/models/rank/dcn/data/download.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os
 import sys
 import io
diff --git a/models/rank/dcn/data/get_slot_data.py b/models/rank/dcn/data/get_slot_data.py
index 77b30296ab9ba99757039f053c6133fd175c2811..96d4448214d6a87092495326646a279657079f45 100755
--- a/models/rank/dcn/data/get_slot_data.py
+++ b/models/rank/dcn/data/get_slot_data.py
@@ -26,8 +26,8 @@ from collections import Counter
 import os
 import paddle.fluid.incubate.data_generator as dg
 
-class TrainReader(dg.MultiSlotDataGenerator):
 
+class TrainReader(dg.MultiSlotDataGenerator):
     def __init__(self, config):
         dg.MultiSlotDataGenerator.__init__(self)
 
@@ -83,11 +83,11 @@ class TrainReader(dg.MultiSlotDataGenerator):
                     if idx == 2 else math.log(1 + float(features[idx])))
         for idx in self.cat_idx_:
             if features[idx] == '' or features[
-                idx] not in self.cat_feat_idx_dict_list[idx - 14]:
+                    idx] not in self.cat_feat_idx_dict_list[idx - 14]:
                 label_feat_list[idx].append(0)
             else:
                 label_feat_list[idx].append(self.cat_feat_idx_dict_list[
-                                                idx - 14][features[idx]])
+                    idx - 14][features[idx]])
         label_feat_list[0].append(int(features[0]))
         return label_feat_list
 
@@ -109,6 +109,7 @@ class TrainReader(dg.MultiSlotDataGenerator):
 
         return data_iter
 
+
 reader = TrainReader("../config.yaml")
 reader.init()
 reader.run_from_stdin()
diff --git a/models/rank/dcn/data/preprocess.py b/models/rank/dcn/data/preprocess.py
index b356607729eedd73854a77449ffda3cc3bb8050f..9a89df10ef42dcfa09faad66f409b21439f340a8 100755
--- a/models/rank/dcn/data/preprocess.py
+++ b/models/rank/dcn/data/preprocess.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from __future__ import print_function, absolute_import, division
 
 import os
diff --git a/models/rank/deepfm/data/download_preprocess.py b/models/rank/deepfm/data/download_preprocess.py
index e8c94cc64728a5e3ae38a29bf419fc90b55df597..7a504b4f88e49d8b4f242d4d6b56f6f168464e5c 100755
--- a/models/rank/deepfm/data/download_preprocess.py
+++ b/models/rank/deepfm/data/download_preprocess.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os
 import shutil
 import sys
diff --git a/models/rank/deepfm/data/get_slot_data.py b/models/rank/deepfm/data/get_slot_data.py
index 59dc33b0d7b6aa1b2087134c9952fc160ca6cd04..6177c990d8ef0c8a1cf922dd9d50c6419cb8c1b7 100755
--- a/models/rank/deepfm/data/get_slot_data.py
+++ b/models/rank/deepfm/data/get_slot_data.py
@@ -19,8 +19,9 @@ try:
     import cPickle as pickle
 except ImportError:
     import pickle
-class TrainReader(dg.MultiSlotDataGenerator):
 
+
+class TrainReader(dg.MultiSlotDataGenerator):
     def __init__(self, config):
         dg.MultiSlotDataGenerator.__init__(self)
 
@@ -44,7 +45,7 @@ class TrainReader(dg.MultiSlotDataGenerator):
         self.categorical_range_ = range(14, 40)
         # load preprocessed feature dict 
         self.feat_dict_name = "aid_data/feat_dict_10.pkl2"
-        self.feat_dict_ = pickle.load(open(self.feat_dict_name, 'rb')) 
+        self.feat_dict_ = pickle.load(open(self.feat_dict_name, 'rb'))
 
     def _process_line(self, line):
         features = line.rstrip('\n').split('\t')
@@ -77,15 +78,18 @@ class TrainReader(dg.MultiSlotDataGenerator):
         def data_iter():
             feat_idx, feat_value, label = self._process_line(line)
             s = ""
-            for i in [('feat_idx', feat_idx), ('feat_value', feat_value), ('label', label)]:
+            for i in [('feat_idx', feat_idx), ('feat_value', feat_value),
+                      ('label', label)]:
                 k = i[0]
                 v = i[1]
                 for j in v:
                     s += " " + k + ":" + str(j)
             print s.strip()
             yield None
+
         return data_iter
 
+
 reader = TrainReader("../config.yaml")
 reader.init()
 reader.run_from_stdin()
diff --git a/models/rank/deepfm/data/preprocess.py b/models/rank/deepfm/data/preprocess.py
index 1fa4a5feae17bde64463d2f05beb3d053284dcda..9da3bdc3d93bfcd0dd98fddc64c870d20feddb38 100755
--- a/models/rank/deepfm/data/preprocess.py
+++ b/models/rank/deepfm/data/preprocess.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os
 import numpy
 from collections import Counter
diff --git a/models/rank/din/data/build_dataset.py b/models/rank/din/data/build_dataset.py
index 34c053ccdb2686c10875740f72f1e0abf3cb4f10..b0ed187800b2f9f44d4dd0d34df204759059ac06 100755
--- a/models/rank/din/data/build_dataset.py
+++ b/models/rank/din/data/build_dataset.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from __future__ import print_function
 import random
 import pickle
diff --git a/models/rank/din/data/convert_pd.py b/models/rank/din/data/convert_pd.py
index d7927c7ef1a9da28732cad9c44be24e72095983a..a66290e1561084a10756ab98c3d70b9a5ac5a6ed 100755
--- a/models/rank/din/data/convert_pd.py
+++ b/models/rank/din/data/convert_pd.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from __future__ import print_function
 import pickle
 import pandas as pd
diff --git a/models/rank/din/data/remap_id.py b/models/rank/din/data/remap_id.py
index b110dac54de8f8d201ede7248d6a2844ac350c90..ee6983d7f0769a58352f61a0a05bbd81c6ccbc13 100755
--- a/models/rank/din/data/remap_id.py
+++ b/models/rank/din/data/remap_id.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from __future__ import print_function
 import random
 import pickle
diff --git a/models/rank/dnn/data/get_slot_data.py b/models/rank/dnn/data/get_slot_data.py
index 30ad9884e5b3c4cd600e8273b9d061bfe1398c9e..f52447d06c297335685a704f688d71aa871328bc 100755
--- a/models/rank/dnn/data/get_slot_data.py
+++ b/models/rank/dnn/data/get_slot_data.py
@@ -32,6 +32,7 @@ class CriteoDataset(dg.MultiSlotDataGenerator):
         """
         Read the data line by line and process it as a dictionary
         """
+
         def reader():
             """
             This function needs to be implemented by the user, based on data format
@@ -57,11 +58,12 @@ class CriteoDataset(dg.MultiSlotDataGenerator):
             feature_name.append("label")
             s = "click:" + str(label[0])
             for i in dense_feature:
-                s += " dense_feature:" + str(i) 
+                s += " dense_feature:" + str(i)
             for i in range(1, 1 + len(categorical_range_)):
-                s += " " + str(i) + ":" + str(sparse_feature[i-1][0])
+                s += " " + str(i) + ":" + str(sparse_feature[i - 1][0])
             print s.strip()
             yield None
+
         return reader
 
 
diff --git a/models/rank/wide_deep/data/data_preparation.py b/models/rank/wide_deep/data/data_preparation.py
index cdd8d4d7817e8312fe76f4038c6554eb557a2ff1..885070096cd3fd084e9695919121f782505b9e77 100644
--- a/models/rank/wide_deep/data/data_preparation.py
+++ b/models/rank/wide_deep/data/data_preparation.py
@@ -1,10 +1,25 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os
 import io
 import args
 import pandas as pd
-from sklearn import  preprocessing
+from sklearn import preprocessing
+
 
-def _clean_file(source_path,target_path):
+def _clean_file(source_path, target_path):
     """makes changes to match the CSV format."""
     with io.open(source_path, 'r') as temp_eval_file:
         with io.open(target_path, 'w') as eval_file:
@@ -17,15 +32,16 @@ def _clean_file(source_path,target_path):
                     line = line[:-1]
                 line += '\n'
                 eval_file.write(line)
-                    
+
+
 def build_model_columns(train_data_path, test_data_path):
     # The column names are from
     # https://www2.1010data.com/documentationcenter/prod/Tutorials/MachineLearningExamples/CensusIncomeDataSet.html
     column_names = [
-    'age', 'workclass', 'fnlwgt', 'education', 'education_num',
-    'marital_status', 'occupation', 'relationship', 'race', 'gender',
-    'capital_gain', 'capital_loss', 'hours_per_week', 'native_country',
-    'income_bracket'
+        'age', 'workclass', 'fnlwgt', 'education', 'education_num',
+        'marital_status', 'occupation', 'relationship', 'race', 'gender',
+        'capital_gain', 'capital_loss', 'hours_per_week', 'native_country',
+        'income_bracket'
     ]
 
     # Load the dataset in Pandas
@@ -44,61 +60,92 @@ def build_model_columns(train_data_path, test_data_path):
 
     # First group of tasks according to the paper
     #label_columns = ['income_50k', 'marital_stat']
-    categorical_columns = ['education','marital_status','relationship','workclass','occupation']
+    categorical_columns = [
+        'education', 'marital_status', 'relationship', 'workclass',
+        'occupation'
+    ]
     for col in categorical_columns:
         label_train = preprocessing.LabelEncoder()
-        train_df[col]= label_train.fit_transform(train_df[col])
+        train_df[col] = label_train.fit_transform(train_df[col])
         label_test = preprocessing.LabelEncoder()
-        test_df[col]= label_test.fit_transform(test_df[col])
-    
-    bins = [18, 25, 30, 35, 40, 45, 50, 55, 60, 65]  
-    train_df['age_buckets'] = pd.cut(train_df['age'].values.tolist(), bins,labels=False)
-    test_df['age_buckets'] = pd.cut(test_df['age'].values.tolist(), bins,labels=False)
-    
-    base_columns = ['education', 'marital_status', 'relationship', 'workclass', 'occupation', 'age_buckets']
-    
-    train_df['education_occupation'] = train_df['education'].astype(str) + '_' + train_df['occupation'].astype(str)    
-    test_df['education_occupation'] = test_df['education'].astype(str) + '_' + test_df['occupation'].astype(str)
-    train_df['age_buckets_education_occupation'] = train_df['age_buckets'].astype(str) + '_' + train_df['education'].astype(str) + '_' + train_df['occupation'].astype(str)
-    test_df['age_buckets_education_occupation'] = test_df['age_buckets'].astype(str) + '_' + test_df['education'].astype(str) + '_' + test_df['occupation'].astype(str)
-    crossed_columns = ['education_occupation','age_buckets_education_occupation']
-    
+        test_df[col] = label_test.fit_transform(test_df[col])
+
+    bins = [18, 25, 30, 35, 40, 45, 50, 55, 60, 65]
+    train_df['age_buckets'] = pd.cut(train_df['age'].values.tolist(),
+                                     bins,
+                                     labels=False)
+    test_df['age_buckets'] = pd.cut(test_df['age'].values.tolist(),
+                                    bins,
+                                    labels=False)
+
+    base_columns = [
+        'education', 'marital_status', 'relationship', 'workclass',
+        'occupation', 'age_buckets'
+    ]
+
+    train_df['education_occupation'] = train_df['education'].astype(
+        str) + '_' + train_df['occupation'].astype(str)
+    test_df['education_occupation'] = test_df['education'].astype(
+        str) + '_' + test_df['occupation'].astype(str)
+    train_df['age_buckets_education_occupation'] = train_df[
+        'age_buckets'].astype(str) + '_' + train_df['education'].astype(
+            str) + '_' + train_df['occupation'].astype(str)
+    test_df['age_buckets_education_occupation'] = test_df[
+        'age_buckets'].astype(str) + '_' + test_df['education'].astype(
+            str) + '_' + test_df['occupation'].astype(str)
+    crossed_columns = [
+        'education_occupation', 'age_buckets_education_occupation'
+    ]
+
     for col in crossed_columns:
         label_train = preprocessing.LabelEncoder()
-        train_df[col]= label_train.fit_transform(train_df[col])
+        train_df[col] = label_train.fit_transform(train_df[col])
         label_test = preprocessing.LabelEncoder()
-        test_df[col]= label_test.fit_transform(test_df[col])
-        
+        test_df[col] = label_test.fit_transform(test_df[col])
+
     wide_columns = base_columns + crossed_columns
-    
-    train_df_temp = pd.get_dummies(train_df[categorical_columns],columns=categorical_columns)
-    test_df_temp = pd.get_dummies(test_df[categorical_columns], columns=categorical_columns)
+
+    train_df_temp = pd.get_dummies(
+        train_df[categorical_columns], columns=categorical_columns)
+    test_df_temp = pd.get_dummies(
+        test_df[categorical_columns], columns=categorical_columns)
     train_df = train_df.join(train_df_temp)
     test_df = test_df.join(test_df_temp)
-    
-    deep_columns = list(train_df_temp.columns)+ ['age','education_num','capital_gain','capital_loss','hours_per_week']
-    
-    train_df['label'] = train_df['income_bracket'].apply(lambda x : 1 if x == '>50K' else 0)
-    test_df['label'] = test_df['income_bracket'].apply(lambda x : 1 if x == '>50K' else 0)
-    
-    with io.open('train_data/columns.txt','w') as f:
-        write_str = str(len(wide_columns)) + '\n' + str(len(deep_columns)) + '\n'
+
+    deep_columns = list(train_df_temp.columns) + [
+        'age', 'education_num', 'capital_gain', 'capital_loss',
+        'hours_per_week'
+    ]
+
+    train_df['label'] = train_df['income_bracket'].apply(
+        lambda x: 1 if x == '>50K' else 0)
+    test_df['label'] = test_df['income_bracket'].apply(
+        lambda x: 1 if x == '>50K' else 0)
+
+    with io.open('train_data/columns.txt', 'w') as f:
+        write_str = str(len(wide_columns)) + '\n' + str(len(
+            deep_columns)) + '\n'
         f.write(write_str)
         f.close()
-    with io.open('test_data/columns.txt','w') as f:
-        write_str = str(len(wide_columns)) + '\n' + str(len(deep_columns)) + '\n'
+    with io.open('test_data/columns.txt', 'w') as f:
+        write_str = str(len(wide_columns)) + '\n' + str(len(
+            deep_columns)) + '\n'
         f.write(write_str)
         f.close()
-    
-    train_df[wide_columns + deep_columns + ['label']].fillna(0).to_csv(train_data_path,index=False)
-    test_df[wide_columns + deep_columns + ['label']].fillna(0).to_csv(test_data_path,index=False)
+
+    train_df[wide_columns + deep_columns + ['label']].fillna(0).to_csv(
+        train_data_path, index=False)
+    test_df[wide_columns + deep_columns + ['label']].fillna(0).to_csv(
+        test_data_path, index=False)
 
 
 def clean_file(train_path, test_path, train_data_path, test_data_path):
     _clean_file(train_path, train_data_path)
     _clean_file(test_path, test_data_path)
 
+
 if __name__ == '__main__':
     args = args.parse_args()
-    clean_file(args.train_path, args.test_path, args.train_data_path, args.test_data_path)  
+    clean_file(args.train_path, args.test_path, args.train_data_path,
+               args.test_data_path)
     build_model_columns(args.train_data_path, args.test_data_path)
diff --git a/models/rank/wide_deep/data/get_slot_data.py b/models/rank/wide_deep/data/get_slot_data.py
index b928ae1267113215aa2b71f8dccffddc0db048fb..831d05665b01649f22a3270ec949ebda2941928d 100755
--- a/models/rank/wide_deep/data/get_slot_data.py
+++ b/models/rank/wide_deep/data/get_slot_data.py
@@ -20,6 +20,7 @@ except ImportError:
     import pickle
 import paddle.fluid.incubate.data_generator as dg
 
+
 class TrainReader(dg.MultiSlotDataGenerator):
     def __init__(self, config):
         dg.MultiSlotDataGenerator.__init__(self)
@@ -50,7 +51,8 @@ class TrainReader(dg.MultiSlotDataGenerator):
             wide_feat, deep_deat, label = self._process_line(line)
 
             s = ""
-            for i in [('wide_input', wide_feat), ('deep_input', deep_deat), ('label', label)]:
+            for i in [('wide_input', wide_feat), ('deep_input', deep_deat),
+                      ('label', label)]:
                 k = i[0]
                 v = i[1]
                 for j in v:
@@ -60,6 +62,7 @@ class TrainReader(dg.MultiSlotDataGenerator):
 
         return data_iter
 
+
 reader = TrainReader("../config.yaml")
 reader.init()
 reader.run_from_stdin()
diff --git a/models/rank/xdeepfm/data/download.py b/models/rank/xdeepfm/data/download.py
index 4afd1ce28ec1ba99006414c6b5116178b8b28142..e46a9ced4a69339f5c5f6c45067d34bbbfa39469 100755
--- a/models/rank/xdeepfm/data/download.py
+++ b/models/rank/xdeepfm/data/download.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os
 import shutil
 import sys
diff --git a/models/rank/xdeepfm/data/get_slot_data.py b/models/rank/xdeepfm/data/get_slot_data.py
index d71444135c2198b426638bc4b2665ec053acb2aa..4426e9647c080dce5debdcdbc3e039ac69a69935 100755
--- a/models/rank/xdeepfm/data/get_slot_data.py
+++ b/models/rank/xdeepfm/data/get_slot_data.py
@@ -21,6 +21,7 @@ except ImportError:
     import pickle
 import paddle.fluid.incubate.data_generator as dg
 
+
 class TrainReader(dg.MultiSlotDataGenerator):
     def __init__(self, config):
         dg.MultiSlotDataGenerator.__init__(self)
@@ -48,7 +49,8 @@ class TrainReader(dg.MultiSlotDataGenerator):
             feat_idx, feat_value, label = self._process_line(line)
 
             s = ""
-            for i in [('feat_idx', feat_idx), ('feat_value', feat_value), ('label', label)]:
+            for i in [('feat_idx', feat_idx), ('feat_value', feat_value),
+                      ('label', label)]:
                 k = i[0]
                 v = i[1]
                 for j in v:
@@ -58,6 +60,7 @@ class TrainReader(dg.MultiSlotDataGenerator):
 
         return data_iter
 
+
 reader = TrainReader("../config.yaml")
 reader.init()
 reader.run_from_stdin()
diff --git a/models/recall/gnn/data_process.sh b/models/recall/gnn/data_process.sh
index 38877b6906ecd65ef190aae5f1dcf5a74cece6d0..fc7ed827e0368c59cab8134d22f78e2200980f18 100755
--- a/models/recall/gnn/data_process.sh
+++ b/models/recall/gnn/data_process.sh
@@ -31,5 +31,3 @@ mv diginetica/train.txt train_data
 
 mkdir test_data
 mv diginetica/test.txt test_data
-
-
diff --git a/models/recall/gnn/raw_data/convert_data.py b/models/recall/gnn/raw_data/convert_data.py
index 2e0e57f1f781f7210c46ef265e1189e99a6f7a96..dfe6bc49fcfca0b98ed5cb0ee9d41832dc5c2205 100755
--- a/models/recall/gnn/raw_data/convert_data.py
+++ b/models/recall/gnn/raw_data/convert_data.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import argparse
 import time
 import pickle
@@ -10,6 +24,7 @@ parser.add_argument(
     help='dataset dir: diginetica/yoochoose1_4/yoochoose1_64/sample')
 opt = parser.parse_args()
 
+
 def process_data(file_type):
     path = os.path.join(opt.data_dir, file_type)
     output_path = os.path.splitext(path)[0] + ".txt"
@@ -23,6 +38,7 @@ def process_data(file_type):
             fout.write(str(data[i][1]))
             fout.write("\n")
 
+
 process_data("train")
 process_data("test")
 
diff --git a/models/recall/gnn/raw_data/download.py b/models/recall/gnn/raw_data/download.py
index 69a1ee20b2d634e9eca47c621dce82ac2d98b5f2..9bebdf1b37e2cd45369c14bb7446c206de8017a0 100755
--- a/models/recall/gnn/raw_data/download.py
+++ b/models/recall/gnn/raw_data/download.py
@@ -1,3 +1,17 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import requests
 import sys
 import time
diff --git a/models/recall/readme.md b/models/recall/readme.md
index 664ced053934d461fb2ed4311a8fd4a1f4d9bd8a..421df1315dc22396f2ff3bb5aec99508435e2c8d 100755
--- a/models/recall/readme.md
+++ b/models/recall/readme.md
@@ -78,4 +78,3 @@ python -m paddlerec.run -m paddlerec.models.recall.youtube_dnn # youtube_dnn
 |       MOVIELENS        |       NCF       |       0.688        |       --          |
 |       --        |       Youtube       |       --        |       --          |
 |       1 Billion Word Language Model Benchmark        |       Word2Vec       |       --         |       0.54          |
-
diff --git a/models/recall/word2vec/prepare_data.sh b/models/recall/word2vec/prepare_data.sh
index 8b78eeedd94f088e206e35729a6b35d349b99039..cfd067350ce1d33112806ab72ca78222381a86f4 100755
--- a/models/recall/word2vec/prepare_data.sh
+++ b/models/recall/word2vec/prepare_data.sh
@@ -35,6 +35,3 @@ wget --no-check-certificate https://paddlerec.bj.bcebos.com/word2vec/test_dir.ta
 tar xzvf test_dir.tar -C raw_data
 mv raw_data/data/test_dir test_data/
 rm -rf raw_data
-
-
-
diff --git a/models/treebased/README.md b/models/treebased/README.md
index 8a8317d17be5148f0652b7944442e1929a7684af..3ceb13b62eba8127aa0394397d141b2abe343a32 100644
--- a/models/treebased/README.md
+++ b/models/treebased/README.md
@@ -24,4 +24,4 @@ TDM是为大规模推荐系统设计的、能承载任意先进模型来高效
    - 如何组网？答：paddle封装了大量的深度学习OP，用户可以根据需求设计自己的网络结构。
    - 训练数据如何组织？答：tdm的训练数据主要为：`user/query emb` 加 `item`的正样本，`item`需要映射到树的某个叶子节点。用户只需准备符合该构成的数据即可。负样本的生成，会基于用户提供的树结构，以及paddle提供的`tdm-sampler op`完成高效的负采样，并自动添加相应的label，参与tdm中深度学习模型的训练。
    - 大规模的数据与模型训练如何实现？答：基于paddle优秀的大规模参数服务器分布式能力，可以实现高效的分布式训练。基于paddle-fleet api，学习门槛极低，且可以灵活的支持增量训练，流式训练等业务需求。
-3. 训练好模型后，可以基于paddle，将检索与打分等流程都融入paddle的组网中，生成inference_model与参数文件，基于PaddlePaddle的预测库或者PaddleLite进行快速部署与高效检索。
\ No newline at end of file
+3. 训练好模型后，可以基于paddle，将检索与打分等流程都融入paddle的组网中，生成inference_model与参数文件，基于PaddlePaddle的预测库或者PaddleLite进行快速部署与高效检索。
diff --git a/models/treebased/tdm/tree/layer_list.txt b/models/treebased/tdm/tree/layer_list.txt
index d8606bc601202390bd9aa54197fac8f34e3c5b59..d1c6c50a10f1b40aa1fbdef7d57bdd600549fb11 100755
--- a/models/treebased/tdm/tree/layer_list.txt
+++ b/models/treebased/tdm/tree/layer_list.txt
@@ -1,4 +1,4 @@
 1,2
 3,4,5,6
 7,8,9,10,11,12,13
-14,15,16,17,18,19,20,21,22,23,24,25
\ No newline at end of file
+14,15,16,17,18,19,20,21,22,23,24,25
diff --git a/run.py b/run.py
index 56999935f21bc1de2b2bc7b4a080da023559174a..c80c647d0c8bab5cd9918429f9bf460b6093335d 100755
--- a/run.py
+++ b/run.py
@@ -26,8 +26,10 @@ from paddlerec.core.utils import util
 engines = {}
 device = ["CPU", "GPU"]
 clusters = ["SINGLE", "LOCAL_CLUSTER", "CLUSTER"]
-engine_choices = ["SINGLE", "LOCAL_CLUSTER", "CLUSTER",
-                  "TDM_SINGLE", "TDM_LOCAL_CLUSTER", "TDM_CLUSTER"]
+engine_choices = [
+    "SINGLE", "LOCAL_CLUSTER", "CLUSTER", "TDM_SINGLE", "TDM_LOCAL_CLUSTER",
+    "TDM_CLUSTER"
+]
 custom_model = ['TDM']
 model_name = ""
 
@@ -66,7 +68,8 @@ def get_engine(args):
     engine = engine.upper()
 
     if engine not in engine_choices:
-        raise ValueError("train.engin can not be chosen in {}".format(engine_choices))
+        raise ValueError("train.engin can not be chosen in {}".format(
+            engine_choices))
 
     print("engines: \n{}".format(engines))
 
@@ -77,8 +80,10 @@ def get_engine(args):
 
 def get_transpiler():
     FNULL = open(os.devnull, 'w')
-    cmd = ["python", "-c",
-           "import paddle.fluid as fluid; fleet_ptr = fluid.core.Fleet(); [fleet_ptr.copy_table_by_feasign(10, 10, [2020, 1010])];"]
+    cmd = [
+        "python", "-c",
+        "import paddle.fluid as fluid; fleet_ptr = fluid.core.Fleet(); [fleet_ptr.copy_table_by_feasign(10, 10, [2020, 1010])];"
+    ]
     proc = subprocess.Popen(cmd, stdout=FNULL, stderr=FNULL, cwd=os.getcwd())
     ret = proc.wait()
     if ret == -11:
@@ -152,7 +157,8 @@ def cluster_engine(args):
         update_workspace(flattens)
 
         envs.set_runtime_environs(flattens)
-        print(envs.pretty_print_envs(flattens, ("Submit Runtime Envs", "Value")))
+        print(envs.pretty_print_envs(flattens, ("Submit Runtime Envs", "Value"
+                                                )))
 
         launch = ClusterEngine(None, args.model)
         return launch
@@ -163,7 +169,8 @@ def cluster_engine(args):
         cluster_envs = {}
         cluster_envs["train.trainer.trainer"] = trainer
         cluster_envs["train.trainer.engine"] = "cluster"
-        cluster_envs["train.trainer.threads"] = envs.get_runtime_environ("CPU_NUM")
+        cluster_envs["train.trainer.threads"] = envs.get_runtime_environ(
+            "CPU_NUM")
         cluster_envs["train.trainer.platform"] = envs.get_platform()
         print("launch {} engine with cluster to with model: {}".format(
             trainer, args.model))
@@ -181,7 +188,8 @@ def cluster_engine(args):
 
 
 def cluster_mpi_engine(args):
-    print("launch cluster engine with cluster to run model: {}".format(args.model))
+    print("launch cluster engine with cluster to run model: {}".format(
+        args.model))
 
     cluster_envs = {}
     cluster_envs["train.trainer.trainer"] = "CtrCodingTrainer"
@@ -209,7 +217,8 @@ def local_cluster_engine(args):
     cluster_envs["train.trainer.platform"] = envs.get_platform()
 
     cluster_envs["CPU_NUM"] = "2"
-    print("launch {} engine with cluster to run model: {}".format(trainer, args.model))
+    print("launch {} engine with cluster to run model: {}".format(trainer,
+                                                                  args.model))
 
     set_runtime_envs(cluster_envs, args.model)
     launch = LocalClusterEngine(cluster_envs, args.model)
@@ -217,10 +226,12 @@ def local_cluster_engine(args):
 
 
 def local_mpi_engine(args):
-    print("launch cluster engine with cluster to run model: {}".format(args.model))
+    print("launch cluster engine with cluster to run model: {}".format(
+        args.model))
     from paddlerec.core.engine.local_mpi import LocalMPIEngine
 
-    print("use 1X1 MPI ClusterTraining at localhost to run model: {}".format(args.model))
+    print("use 1X1 MPI ClusterTraining at localhost to run model: {}".format(
+        args.model))
 
     mpi = util.run_which("mpirun")
     if not mpi:
diff --git a/setup.py b/setup.py
index c655c37576e310fac825bd1cc01dfca5d051d18c..31bb34f03187dc9ab29c4cc5c75c559540ca8269 100644
--- a/setup.py
+++ b/setup.py
@@ -11,7 +11,6 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
 """
 setup for paddle-rec.
 """
@@ -22,11 +21,7 @@ from setuptools import setup, find_packages
 import shutil
 import tempfile
 
-
-requires = [
-    "paddlepaddle == 1.7.2",
-    "pyyaml >= 5.1.1"
-]
+requires = ["paddlepaddle == 1.7.2", "pyyaml >= 5.1.1"]
 
 about = {}
 about["__title__"] = "paddle-rec"
@@ -48,18 +43,27 @@ def build(dirname):
     package_dir = os.path.dirname(os.path.abspath(__file__))
     run_cmd("cp -r {}/* {}".format(package_dir, dirname))
     run_cmd("mkdir {}".format(os.path.join(dirname, "paddlerec")))
-    run_cmd("mv {} {}".format(os.path.join(dirname, "core"), os.path.join(dirname, "paddlerec")))
-    run_cmd("mv {} {}".format(os.path.join(dirname, "doc"), os.path.join(dirname, "paddlerec")))
-    run_cmd("mv {} {}".format(os.path.join(dirname, "models"), os.path.join(dirname, "paddlerec")))
-    run_cmd("mv {} {}".format(os.path.join(dirname, "tests"), os.path.join(dirname, "paddlerec")))
-    run_cmd("mv {} {}".format(os.path.join(dirname, "tools"), os.path.join(dirname, "paddlerec")))
-    run_cmd("mv {} {}".format(os.path.join(dirname, "*.py"), os.path.join(dirname, "paddlerec")))
+    run_cmd("mv {} {}".format(
+        os.path.join(dirname, "core"), os.path.join(dirname, "paddlerec")))
+    run_cmd("mv {} {}".format(
+        os.path.join(dirname, "doc"), os.path.join(dirname, "paddlerec")))
+    run_cmd("mv {} {}".format(
+        os.path.join(dirname, "models"), os.path.join(dirname, "paddlerec")))
+    run_cmd("mv {} {}".format(
+        os.path.join(dirname, "tests"), os.path.join(dirname, "paddlerec")))
+    run_cmd("mv {} {}".format(
+        os.path.join(dirname, "tools"), os.path.join(dirname, "paddlerec")))
+    run_cmd("mv {} {}".format(
+        os.path.join(dirname, "*.py"), os.path.join(dirname, "paddlerec")))
 
     packages = find_packages(dirname, include=('paddlerec.*'))
     package_dir = {'': dirname}
     package_data = {}
 
-    models_copy = ['data/*.txt', 'data/*/*.txt', '*.yaml', '*.sh', 'tree/*.npy', 'tree/*.txt']
+    models_copy = [
+        'data/*.txt', 'data/*/*.txt', '*.yaml', '*.sh', 'tree/*.npy',
+        'tree/*.txt'
+    ]
     engine_copy = ['*/*.sh']
     for package in packages:
         if package.startswith("paddlerec.models."):
@@ -80,8 +84,7 @@ def build(dirname):
         package_data=package_data,
         python_requires=">=2.7",
         install_requires=requires,
-        zip_safe=False
-    )
+        zip_safe=False)
 
 
 dirname = tempfile.mkdtemp()