Deprecate --disable-log-requests and replace with --enable-log-requests (#21739)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
@@ -18,7 +18,7 @@ from typing import (TYPE_CHECKING, Annotated, Any, Callable, Dict, List,
|
||||
import regex as re
|
||||
import torch
|
||||
from pydantic import TypeAdapter, ValidationError
|
||||
from typing_extensions import TypeIs
|
||||
from typing_extensions import TypeIs, deprecated
|
||||
|
||||
import vllm.envs as envs
|
||||
from vllm.config import (BlockSize, CacheConfig, CacheDType, CompilationConfig,
|
||||
@@ -1704,7 +1704,23 @@ class EngineArgs:
|
||||
@dataclass
|
||||
class AsyncEngineArgs(EngineArgs):
|
||||
"""Arguments for asynchronous vLLM engine."""
|
||||
disable_log_requests: bool = False
|
||||
enable_log_requests: bool = False
|
||||
|
||||
@property
|
||||
@deprecated(
|
||||
"`disable_log_requests` is deprecated and has been replaced with "
|
||||
"`enable_log_requests`. This will be removed in v0.12.0. Please use "
|
||||
"`enable_log_requests` instead.")
|
||||
def disable_log_requests(self) -> bool:
|
||||
return not self.enable_log_requests
|
||||
|
||||
@disable_log_requests.setter
|
||||
@deprecated(
|
||||
"`disable_log_requests` is deprecated and has been replaced with "
|
||||
"`enable_log_requests`. This will be removed in v0.12.0. Please use "
|
||||
"`enable_log_requests` instead.")
|
||||
def disable_log_requests(self, value: bool):
|
||||
self.enable_log_requests = not value
|
||||
|
||||
@staticmethod
|
||||
def add_cli_args(parser: FlexibleArgumentParser,
|
||||
@@ -1715,9 +1731,15 @@ class AsyncEngineArgs(EngineArgs):
|
||||
load_general_plugins()
|
||||
if not async_args_only:
|
||||
parser = EngineArgs.add_cli_args(parser)
|
||||
parser.add_argument('--enable-log-requests',
|
||||
action=argparse.BooleanOptionalAction,
|
||||
default=AsyncEngineArgs.enable_log_requests,
|
||||
help='Enable logging requests.')
|
||||
parser.add_argument('--disable-log-requests',
|
||||
action='store_true',
|
||||
help='Disable logging requests.')
|
||||
action=argparse.BooleanOptionalAction,
|
||||
default=not AsyncEngineArgs.enable_log_requests,
|
||||
help='[DEPRECATED] Disable logging requests.',
|
||||
deprecated=True)
|
||||
current_platform.pre_register_and_update(parser)
|
||||
return parser
|
||||
|
||||
|
||||
@@ -30,7 +30,7 @@ from vllm.sampling_params import SamplingParams
|
||||
from vllm.sequence import ExecuteModelRequest
|
||||
from vllm.transformers_utils.tokenizer import AnyTokenizer
|
||||
from vllm.usage.usage_lib import UsageContext
|
||||
from vllm.utils import Device, weak_bind
|
||||
from vllm.utils import Device, deprecate_kwargs, weak_bind
|
||||
|
||||
logger = init_logger(__name__)
|
||||
ENGINE_ITERATION_TIMEOUT_S = envs.VLLM_ENGINE_ITERATION_TIMEOUT_S
|
||||
@@ -554,14 +554,20 @@ class AsyncLLMEngine(EngineClient):
|
||||
return LLMEngine._get_executor_cls(engine_config)
|
||||
|
||||
@classmethod
|
||||
@deprecate_kwargs(
|
||||
"disable_log_requests",
|
||||
additional_message=("This argument will have no effect. "
|
||||
"Use `enable_log_requests` instead."),
|
||||
)
|
||||
def from_vllm_config(
|
||||
cls,
|
||||
vllm_config: VllmConfig,
|
||||
start_engine_loop: bool = True,
|
||||
usage_context: UsageContext = UsageContext.ENGINE_CONTEXT,
|
||||
stat_loggers: Optional[dict[str, StatLoggerBase]] = None,
|
||||
disable_log_requests: bool = False,
|
||||
disable_log_stats: bool = False,
|
||||
cls,
|
||||
vllm_config: VllmConfig,
|
||||
start_engine_loop: bool = True,
|
||||
usage_context: UsageContext = UsageContext.ENGINE_CONTEXT,
|
||||
stat_loggers: Optional[dict[str, StatLoggerBase]] = None,
|
||||
enable_log_requests: bool = False,
|
||||
disable_log_stats: bool = False,
|
||||
disable_log_requests: bool = True, # Deprecated, will be removed
|
||||
) -> "AsyncLLMEngine":
|
||||
"""Create an AsyncLLMEngine from the EngineArgs."""
|
||||
|
||||
@@ -569,7 +575,7 @@ class AsyncLLMEngine(EngineClient):
|
||||
vllm_config=vllm_config,
|
||||
executor_class=cls._get_executor_cls(vllm_config),
|
||||
start_engine_loop=start_engine_loop,
|
||||
log_requests=not disable_log_requests,
|
||||
log_requests=enable_log_requests,
|
||||
log_stats=not disable_log_stats,
|
||||
usage_context=usage_context,
|
||||
stat_loggers=stat_loggers,
|
||||
@@ -598,7 +604,7 @@ class AsyncLLMEngine(EngineClient):
|
||||
usage_context=usage_context,
|
||||
stat_loggers=stat_loggers,
|
||||
disable_log_stats=engine_args.disable_log_stats,
|
||||
disable_log_requests=engine_args.disable_log_requests,
|
||||
enable_log_requests=engine_args.enable_log_requests,
|
||||
)
|
||||
|
||||
@property
|
||||
|
||||
@@ -34,6 +34,7 @@ from vllm.outputs import RequestOutput
|
||||
from vllm.transformers_utils.config import (
|
||||
maybe_register_config_serialize_by_value)
|
||||
from vllm.usage.usage_lib import UsageContext
|
||||
from vllm.utils import deprecate_kwargs
|
||||
from vllm.worker.model_runner_base import InputProcessingError
|
||||
|
||||
logger = init_logger(__name__)
|
||||
@@ -120,10 +121,20 @@ class MQLLMEngine:
|
||||
return ENGINE_DEAD_ERROR()
|
||||
|
||||
@classmethod
|
||||
def from_vllm_config(cls, vllm_config: VllmConfig,
|
||||
usage_context: UsageContext,
|
||||
disable_log_requests: bool, disable_log_stats: bool,
|
||||
ipc_path: str) -> "MQLLMEngine":
|
||||
@deprecate_kwargs(
|
||||
"disable_log_requests",
|
||||
additional_message=("This argument will have no effect. "
|
||||
"Use `enable_log_requests` instead."),
|
||||
)
|
||||
def from_vllm_config(
|
||||
cls,
|
||||
vllm_config: VllmConfig,
|
||||
usage_context: UsageContext,
|
||||
enable_log_requests: bool,
|
||||
disable_log_stats: bool,
|
||||
ipc_path: str,
|
||||
disable_log_requests: bool = True, # Deprecated, will be removed
|
||||
) -> "MQLLMEngine":
|
||||
# Setup plugins for each process
|
||||
from vllm.plugins import load_general_plugins
|
||||
load_general_plugins()
|
||||
@@ -136,7 +147,7 @@ class MQLLMEngine:
|
||||
ipc_path=ipc_path,
|
||||
usage_context=usage_context,
|
||||
use_async_sockets=use_async_sockets,
|
||||
log_requests=(not disable_log_requests),
|
||||
log_requests=enable_log_requests,
|
||||
log_stats=(not disable_log_stats),
|
||||
)
|
||||
|
||||
@@ -150,7 +161,7 @@ class MQLLMEngine:
|
||||
ipc_path=ipc_path,
|
||||
vllm_config=vllm_config,
|
||||
usage_context=usage_context,
|
||||
disable_log_requests=engine_args.disable_log_requests,
|
||||
enable_log_requests=engine_args.enable_log_requests,
|
||||
disable_log_stats=engine_args.disable_log_stats,
|
||||
)
|
||||
|
||||
@@ -436,7 +447,7 @@ def signal_handler(*_) -> None:
|
||||
|
||||
def run_mp_engine(vllm_config: VllmConfig, usage_context: UsageContext,
|
||||
ipc_path: str, disable_log_stats: bool,
|
||||
disable_log_requests: bool, engine_alive):
|
||||
enable_log_requests: bool, engine_alive):
|
||||
try:
|
||||
# Ensure we can serialize transformer config before spawning
|
||||
maybe_register_config_serialize_by_value()
|
||||
@@ -445,7 +456,7 @@ def run_mp_engine(vllm_config: VllmConfig, usage_context: UsageContext,
|
||||
vllm_config=vllm_config,
|
||||
usage_context=usage_context,
|
||||
disable_log_stats=disable_log_stats,
|
||||
disable_log_requests=disable_log_requests,
|
||||
enable_log_requests=enable_log_requests,
|
||||
ipc_path=ipc_path)
|
||||
|
||||
signal.signal(signal.SIGTERM, signal_handler)
|
||||
|
||||
Reference in New Issue
Block a user