xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-17 12:35:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Register serializers for V0 MQ Engine (#15009)

Browse Source 
Signed-off-by: simon-mo <simon.mo@hey.com>
This commit is contained in:
Simon Mo 2025-03-18 06:14:47 -07:00 committed by GitHub
parent ab656f2c2f
commit 3b457143d2
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 13 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
vllm/engine/multiprocessing/engine.py
View File

@ -29,6 +29,8 @@ from vllm.engine.multiprocessing import (ENGINE_DEAD_ERROR, IPC_DATA_EXT,
# yapf: enable # yapf: enable
from vllm.logger import init_logger from vllm.logger import init_logger
from vllm.outputs import RequestOutput 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.usage.usage_lib import UsageContext
from vllm.worker.model_runner_base import InputProcessingError from vllm.worker.model_runner_base import InputProcessingError
@ -42,12 +44,12 @@ class MQLLMEngine:
"""A multiprocessing wrapper for :class:`LLMEngine`. """A multiprocessing wrapper for :class:`LLMEngine`.
This class is used to wrap the :class:`LLMEngine` class to enable use This class is used to wrap the :class:`LLMEngine` class to enable use
in concurrnet manner. It runs a background loop and uses zeromq to in concurrnet manner. It runs a background loop and uses zeromq to
receive new requests and stream outputs incrementally via ipc. receive new requests and stream outputs incrementally via ipc.
The :class:`LLMEngine` generate or encode process is kicked off when a new The :class:`LLMEngine` generate or encode process is kicked off when a new
RPCProcessRequest is received by the input_socket. RPCProcessRequest is received by the input_socket.
The self.engine_loop checks the input_socket for new requests, The self.engine_loop checks the input_socket for new requests,
adds them to the LLMEngine if there are any, calls the internal adds them to the LLMEngine if there are any, calls the internal
:class:`LLMEngine.step()`, and sends the RequestOutputs back over :class:`LLMEngine.step()`, and sends the RequestOutputs back over
@ -428,6 +430,9 @@ def run_mp_engine(vllm_config: VllmConfig, usage_context: UsageContext,
ipc_path: str, disable_log_stats: bool, ipc_path: str, disable_log_stats: bool,
disable_log_requests: bool, engine_alive): disable_log_requests: bool, engine_alive):
try: try:
# Ensure we can serialize transformer config before spawning
maybe_register_config_serialize_by_value()
engine = MQLLMEngine.from_vllm_config( engine = MQLLMEngine.from_vllm_config(
vllm_config=vllm_config, vllm_config=vllm_config,
usage_context=usage_context, usage_context=usage_context,

5
vllm/entrypoints/openai/api_server.py
View File

@ -82,6 +82,8 @@ from vllm.entrypoints.openai.serving_transcription import (
from vllm.entrypoints.openai.tool_parsers import ToolParserManager from vllm.entrypoints.openai.tool_parsers import ToolParserManager
from vllm.entrypoints.utils import load_aware_call, with_cancellation from vllm.entrypoints.utils import load_aware_call, with_cancellation
from vllm.logger import init_logger from vllm.logger import init_logger
from vllm.transformers_utils.config import (
maybe_register_config_serialize_by_value)
from vllm.usage.usage_lib import UsageContext from vllm.usage.usage_lib import UsageContext
from vllm.utils import (FlexibleArgumentParser, get_open_zmq_ipc_path, from vllm.utils import (FlexibleArgumentParser, get_open_zmq_ipc_path,
is_valid_ipv6_address, set_ulimit) is_valid_ipv6_address, set_ulimit)
@ -221,6 +223,9 @@ async def build_async_engine_client_from_engine_args(
# so we need to spawn a new process # so we need to spawn a new process
context = multiprocessing.get_context("spawn") context = multiprocessing.get_context("spawn")
# Ensure we can serialize transformer config before spawning
maybe_register_config_serialize_by_value()
# The Process can raise an exception during startup, which may # The Process can raise an exception during startup, which may
# not actually result in an exitcode being reported. As a result # not actually result in an exitcode being reported. As a result
# we use a shared variable to communicate the information. # we use a shared variable to communicate the information.