Do not overwrite root logger

python/paddle/distributed/auto_parallel/operators/dist_flash_attn.py

@@ -16,7 +16,7 @@
 
 from ...utils.log_utils import get_logger
 
-_logger = get_logger(logging.INFO)
+_logger = get_logger(logging.INFO, __name__)
 from ..random import determinate_rng, is_enable_auto_rand_ctrl
 from .common import (
     DistributedOperatorImplContainer,

python/paddle/distributed/auto_parallel/random.py

@@ -22,7 +22,7 @@
 from .process_mesh import retrieve_unique_id_for_process_mesh
 from .static.utils import _get_idx_in_axis
 
-_logger = get_logger(logging.INFO)
+_logger = get_logger(logging.INFO, __name__)
 
 _rng_name_to_seed = {}
 _rng_name_to_states = {}

python/paddle/distributed/auto_parallel/static/cluster.py

@@ -1230,7 +1230,7 @@ def __repr__(self):
         return self.__str__()
 
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 def get_default_cluster(json_config=None, auto_config=None):

python/paddle/distributed/auto_parallel/static/converter.py

@@ -44,7 +44,7 @@ def __init__(self, tensors_dict, pre_strategy, cur_strategy):
         self._tensors_dict = self._check_tensor_dict(tensors_dict)
         self._pre_strategy = self._check_pre_strategy(pre_strategy)
         self._cur_strategy = self._check_cur_strategy(cur_strategy)
-        self._logger = get_logger(logging.INFO)
+        self._logger = get_logger(logging.INFO, __name__)
 
     def _check_tensor_dict(self, tensors_dict):
         if not tensors_dict:

python/paddle/distributed/auto_parallel/static/dist_saver.py

@@ -52,7 +52,7 @@ def _process_path(path):
 
 class DistributedSaver:
     def __init__(self):
-        self._logger = get_logger(logging.INFO)
+        self._logger = get_logger(logging.INFO, __name__)
 
     def save(self, path, serial_program, dist_main_program, dist_context):
         def _save_state(program, path, mode="param"):

python/paddle/distributed/auto_parallel/static/engine.py

@@ -228,7 +228,7 @@ def __init__(
             )
         self._strategy = strategy or Strategy()
 
-        self._logger = get_logger(logging.INFO)
+        self._logger = get_logger(logging.INFO, __name__)
 
         # for compute cost
         # TODO: remove _fwd_main_progs and _orig_optimizer and _pir_main_progs

python/paddle/distributed/auto_parallel/static/parallelizer_v2.py

@@ -51,7 +51,7 @@ def __init__(self, mode, completer, dist_context):
         assert self._dist_context._is_initialized
         self._pass_context = self._dist_context.pass_context
         self._strategy = self._dist_context.strategy
-        self._logger = get_logger(logging.INFO)
+        self._logger = get_logger(logging.INFO, __name__)
 
     @property
     def is_train(self):

python/paddle/distributed/auto_parallel/static/planner_v2.py

@@ -73,7 +73,7 @@ def completer(self):
         return self._completer
 
     def plan(self):
-        logger = get_logger(logging.INFO)
+        logger = get_logger(logging.INFO, __name__)
         path = None
         if self._dist_context._json_config:
             try:

python/paddle/distributed/auto_parallel/static/tuner/rule_based_tuner.py

@@ -1056,7 +1056,7 @@ def __init__(self, dist_context, mode="train", level="o1"):
         self._mode = mode
         assert level in ["o1", "o2"]
         self._level = level
-        self._logger = get_logger(logging.INFO)
+        self._logger = get_logger(logging.INFO, __name__)
         self._use_dp = False
 
         # forward sub program

python/paddle/distributed/fleet/meta_parallel/sharding/group_sharded_stage2.py

@@ -37,7 +37,7 @@
 from .group_sharded_storage import GradStorage
 from .group_sharded_utils import Type, device_guard
 
-logger_ = get_logger(logging.WARNING)
+logger_ = get_logger(logging.WARNING, __name__)
 
 
 def _trainable(param):

python/paddle/distributed/passes/auto_parallel_recompute.py

@@ -41,7 +41,7 @@
 from ..utils.log_utils import get_logger
 from .pass_base import PassBase, register_pass
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 class RecomputeState(ProgramStats):

python/paddle/distributed/passes/pipeline_scheduler_pass/pipeline_1f1b.py

@@ -29,7 +29,7 @@
 )
 from .pipeline_pass_base import PipelinePassBase
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 @register_pass("pipeline_scheduler_1F1B")

python/paddle/distributed/passes/pipeline_scheduler_pass/pipeline_eager_1f1b.py

@@ -20,7 +20,7 @@
 from ..pass_base import register_pass
 from .pipeline_pass_base import PipelinePassBase
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 @register_pass("pipeline_scheduler_Eager1F1B")

python/paddle/distributed/passes/pipeline_scheduler_pass/pipeline_fthenb.py

@@ -23,7 +23,7 @@
 )
 from .pipeline_pass_base import PipelinePassBase
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 @register_pass("pipeline_scheduler_FThenB")

python/paddle/distributed/passes/pipeline_scheduler_pass/pipeline_pass_base.py

@@ -23,7 +23,7 @@
     set_skip_gc_vars,
 )
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 class PipelinePassBase(PassBase):

python/paddle/distributed/passes/pipeline_scheduler_pass/pipeline_vpp.py

@@ -31,7 +31,7 @@
 )
 from .pipeline_pass_base import PipelinePassBase
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 @register_pass("pipeline_scheduler_VPP")

python/paddle/distributed/passes/pipeline_scheduler_pass/pipeline_zero_bubble.py

@@ -21,7 +21,7 @@
 from ..pass_base import register_pass
 from .pipeline_pass_base import PipelinePassBase
 
-logger = get_logger(logging.INFO)
+logger = get_logger(logging.INFO, __name__)
 
 
 class PipelineZeroBubbleBase(PipelinePassBase):

python/paddle/distributed/sharding/group_sharded.py

@@ -44,7 +44,7 @@
     from paddle.distributed.communication.group import Group
     from paddle.nn import Layer
 
-logger_ = get_logger(logging.WARNING)
+logger_ = get_logger(logging.WARNING, __name__)
 
 
 def group_sharded_parallel(

python/paddle/distributed/utils/launch_utils.py

@@ -26,7 +26,7 @@
 
 from ..utils.log_utils import get_logger
 
-logger = get_logger("INFO", "root")
+logger = get_logger("INFO", __name__)
 
 
 def get_cluster_from_args(args, selected_gpus):

python/paddle/distributed/utils/log_utils.py

@@ -15,7 +15,7 @@
 import logging
 
 
-def get_logger(log_level, name="root"):
+def get_logger(log_level, name):
     logger = logging.getLogger(name)
 
     # Avoid printing multiple logs

python/paddle/distributed/utils/process_utils.py

@@ -19,7 +19,7 @@
 import paddle
 from paddle.distributed.utils.log_utils import get_logger
 
-logger = get_logger("INFO", "root")
+logger = get_logger("INFO", __name__)
 
 SUCCESS_CODE = 0
 FAIL_CODE = 1