fix and change all logger using get_logger (#1064)

* fix logger

* Merge the main branch for logger

* Fix some unchanged logging to logger

benchmark/profile_hf_generation.py

@@ -75,8 +75,9 @@
 from transformers import AutoModelForCausalLM, GenerationConfig
 
 from lmdeploy.pytorch.accel import LoadNoInit
+from lmdeploy.utils import get_logger
 
-logger = logging.getLogger(__file__)
+logger = get_logger(__file__)
 logger.setLevel(logging.DEBUG)
 info = logger.info
 warning = logger.warning

lmdeploy/legacy/pytorch/adapters/__init__.py

@@ -1,14 +1,14 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 
-import logging
-
 import torch.nn as nn
 
+from lmdeploy.utils import get_logger
+
 from .base import BasicAdapter, BasicAdapterFast
 from .internlm import InternLMAdapter
 from .llama2 import Llama2Adapter
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 def _get_default_adapter(tokenizer):

lmdeploy/legacy/pytorch/adapters/base.py

@@ -1,13 +1,14 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 """Basic adapter suitable for general HuggingFace models."""
 
-import logging
 import re
 
 from transformers import (PreTrainedTokenizer, PreTrainedTokenizerBase,
                           PreTrainedTokenizerFast)
 
-logger = logging.getLogger(__name__)
+from lmdeploy.utils import get_logger
+
+logger = get_logger(__name__)
 
 
 class BaseAdapter:

lmdeploy/legacy/pytorch/adapters/internlm.py

@@ -1,14 +1,15 @@
 # Copyright (c) OpenMMLab. All rights reserved.
-import logging
 import re
 
 import torch
 from transformers import (PreTrainedTokenizerFast, StoppingCriteria,
                           StoppingCriteriaList)
 
+from lmdeploy.utils import get_logger
+
 from .base import BaseAdapter
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 class InternLMStoppingCriteria(StoppingCriteria):

lmdeploy/legacy/pytorch/adapters/llama2.py

@@ -1,12 +1,13 @@
 # Copyright (c) OpenMMLab. All rights reserved.
-import logging
 import re
 
 from transformers import PreTrainedTokenizerFast
 
+from lmdeploy.utils import get_logger
+
 from .base import BasicAdapterFast
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 B_INST, E_INST = '[INST]', '[/INST]'
 B_SYS, E_SYS = '<<SYS>>\n', '\n<</SYS>>\n\n'

lmdeploy/legacy/pytorch/chat.py

@@ -54,30 +54,34 @@
 import torch
 from transformers import GenerationConfig, PreTrainedModel
 
+from lmdeploy.utils import get_logger
+
 from .adapters import init_adapter
 from .dist import get_local_rank, get_rank, get_world_size
 from .model import accel_model, init_model
 from .session import BasicSessionManagerWithHistory
 from .utils import BasicStreamer, TerminalIO, control
 
-logger = logging.getLogger(__name__)
-
 
 def set_logging(log_file: str, debug: bool):
     torch.set_printoptions(linewidth=120)
     level = logging.DEBUG if debug else logging.INFO
     log_file = log_file or 'chat.log'
     if r := get_rank() != 0:
         log_file = log_file + f'.{r}'
-    logging.basicConfig(level=level,
-                        format=('%(filename)s: '
-                                '%(levelname)s: '
-                                '%(funcName)s(): '
-                                '%(lineno)d:\t'
-                                '%(message)s'),
-                        filename=log_file,
-                        filemode='w')
+    format = '%(filename)s:   \
+              %(levelname)s:  \
+              %(funcName)s(): \
+              %(lineno)d:\t   \
+              %(message)s'
+
+    logger = get_logger(__name__,
+                        log_file=log_file,
+                        log_level=level,
+                        file_mode='w',
+                        log_formatter=format)
     print(f'Worker {get_rank()} logging to {log_file}')
+    return logger
 
 
 def main(
@@ -120,7 +124,7 @@ def main(
             based on `LlamaforCausalLM` class, this argument is required.
             Currently, only "llama1" is acceptable for llama1 models.
     """  # noqa: E501
-    set_logging(log_file, debug)
+    logger = set_logging(log_file, debug)
 
     # workers should sync in sampling
     torch.manual_seed(seed)

lmdeploy/legacy/pytorch/decode.py

@@ -1,6 +1,5 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 import argparse
-import logging
 import queue
 import warnings
 from typing import List, Optional
@@ -12,6 +11,8 @@
 from transformers import (AutoTokenizer, PreTrainedModel,
                           PreTrainedTokenizerBase)
 
+from lmdeploy.utils import get_logger
+
 from .model import accel_model, init_model
 
 
@@ -371,7 +372,7 @@ def __del__(self):
     test_path = args.test_path
     prompts = args.prompts
 
-    logger = logging.getLogger(__name__)
+    logger = get_logger(__name__)
     # logging.basicConfig(level=logging.DEBUG)
 
     # Use test file preferentially

lmdeploy/legacy/pytorch/model.py

@@ -1,15 +1,16 @@
 # Copyright (c) OpenMMLab. All rights reserved.
-import logging
 import time
 import warnings
 from typing import Optional
 
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
+from lmdeploy.utils import get_logger
+
 from .dist import get_local_rank
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 class LoadWoInit:

lmdeploy/legacy/pytorch/session.py

@@ -1,10 +1,11 @@
 # Copyright (c) OpenMMLab. All rights reserved.
-import logging
 
 import torch
 from transformers.generation.utils import ModelOutput
 
-logger = logging.getLogger(__name__)
+from lmdeploy.utils import get_logger
+
+logger = get_logger(__name__)
 
 
 class BasicSessionManager:

lmdeploy/legacy/pytorch/utils.py

@@ -1,17 +1,17 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 
-import logging
-
 from transformers.generation.streamers import BaseStreamer
 
+from lmdeploy.utils import get_logger
+
 from .dist import get_rank, master_only, master_only_and_broadcast_general
 
 try:
     import readline  # To support command line history # noqa: F401
 except ImportError:  # readline not available
     pass
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 class TerminalIO:

lmdeploy/pytorch/modeling/modeling_baichuan.py

@@ -27,13 +27,13 @@
 from transformers.activations import ACT2FN
 from transformers.modeling_outputs import (BaseModelOutputWithPast,
                                            CausalLMOutputWithPast)
-from transformers.utils import logging
 
 from lmdeploy.pytorch.modeling.convert_to_qmodules import convert_to_qmodules
+from lmdeploy.utils import get_logger
 
 from .configuration_baichuan import BaiChuanConfig
 
-logger = logging.get_logger(__name__)
+logger = get_logger(__name__)
 
 
 # Copied from transformers.models.bart.modeling_bart._make_causal_mask

lmdeploy/pytorch/modeling/modeling_internlm.py

@@ -33,14 +33,15 @@
                                            SequenceClassifierOutputWithPast)
 from transformers.modeling_utils import PreTrainedModel
 from transformers.utils import (add_start_docstrings,
-                                add_start_docstrings_to_model_forward, logging,
+                                add_start_docstrings_to_model_forward,
                                 replace_return_docstrings)
 
 from lmdeploy.pytorch.modeling.convert_to_qmodules import convert_to_qmodules
+from lmdeploy.utils import get_logger
 
 from .configuration_internlm import InternLMConfig
 
-logger = logging.get_logger(__name__)
+logger = get_logger(__name__)
 
 _CONFIG_FOR_DOC = 'InternLMConfig'
 

lmdeploy/pytorch/modeling/modeling_internlm2.py

@@ -34,10 +34,11 @@
                                            SequenceClassifierOutputWithPast)
 from transformers.modeling_utils import PreTrainedModel
 from transformers.utils import (add_start_docstrings,
-                                add_start_docstrings_to_model_forward, logging,
+                                add_start_docstrings_to_model_forward,
                                 replace_return_docstrings)
 
 from lmdeploy.pytorch.modeling.convert_to_qmodules import convert_to_qmodules
+from lmdeploy.utils import get_logger
 
 try:
     from transformers.generation.streamers import BaseStreamer
@@ -46,7 +47,7 @@
 
 from .configuration_internlm import InternLMConfig as InternLM2Config
 
-logger = logging.get_logger(__name__)
+logger = get_logger(__name__)
 
 _CONFIG_FOR_DOC = 'InternLM2Config'
 

lmdeploy/pytorch/modeling/modeling_llama.py

@@ -32,12 +32,13 @@
 from transformers.modeling_utils import PreTrainedModel
 from transformers.models.llama.configuration_llama import LlamaConfig
 from transformers.utils import (add_start_docstrings,
-                                add_start_docstrings_to_model_forward, logging,
+                                add_start_docstrings_to_model_forward,
                                 replace_return_docstrings)
 
 from lmdeploy.pytorch.modeling.convert_to_qmodules import convert_to_qmodules
+from lmdeploy.utils import get_logger
 
-logger = logging.get_logger(__name__)
+logger = get_logger(__name__)
 
 _CONFIG_FOR_DOC = 'LlamaConfig'
 

lmdeploy/pytorch/models/falcon.py

@@ -3,7 +3,6 @@
 # https://huggingface.co/tiiuae/falcon-7b-instruct
 # https://github.com/huggingface/transformers/blob/v4.33-release/src/transformers/models/falcon/modeling_falcon.py  # noqa
 
-import logging
 from typing import Optional, Tuple, Union
 
 import torch
@@ -20,8 +19,6 @@
 from ..kernels import (alibi_paged_attention_fwd, fill_kv_cache,
                        paged_attention_fwd)
 
-logger = logging.getLogger()
-
 
 # rotary pos emb helpers
 # (torch.jit.script does not seem to support staticmethod...)

lmdeploy/serve/qos_engine/inner_group_schd.py

@@ -1,8 +1,9 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 import collections
-import logging
 
-logger = logging.getLogger(__name__)
+from lmdeploy.utils import get_logger
+
+logger = get_logger(__name__)
 
 
 class UserRequestQueue:

lmdeploy/serve/qos_engine/qos_engine.py

@@ -1,7 +1,6 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 import asyncio
 import json
-import logging
 import threading
 import time
 from typing import List
@@ -11,8 +10,9 @@
                                             GenerateRequestQos)
 from lmdeploy.serve.qos_engine.inner_group_schd import UserRequestQueue
 from lmdeploy.serve.qos_engine.usage_stats import UsageStats
+from lmdeploy.utils import get_logger
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 class QosConfig: