biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V0 Deprecation] Remove AsyncLLMEngine (#25025)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon
2025-09-18 11:07:42 -07:00
committed by GitHub
parent 505805b645
commit e19bce40a1
11 changed files with 76 additions and 1967 deletions
Download Patch File Download Diff File Expand all files Collapse all files

65
vllm/entrypoints/openai/api_server.py
View File

@@ -38,7 +38,6 @@ from typing_extensions import assert_never
import vllm.envs as envs
from vllm.config import VllmConfig
from vllm.engine.arg_utils import AsyncEngineArgs
from vllm.engine.async_llm_engine import AsyncLLMEngine # type: ignore
from vllm.engine.protocol import EngineClient
from vllm.entrypoints.chat_utils import (load_chat_template,
resolve_hf_chat_template,
@@ -201,50 +200,34 @@ async def build_async_engine_client_from_engine_args(
vllm_config = engine_args.create_engine_config(usage_context=usage_context)
# V1 AsyncLLM.
if envs.VLLM_USE_V1:
if disable_frontend_multiprocessing:
logger.warning(
"V1 is enabled, but got --disable-frontend-multiprocessing. "
"To disable frontend multiprocessing, set VLLM_USE_V1=0.")
assert envs.VLLM_USE_V1
from vllm.v1.engine.async_llm import AsyncLLM
async_llm: Optional[AsyncLLM] = None
client_count = client_config.pop(
"client_count") if client_config else 1
client_index = client_config.pop(
"client_index") if client_config else 0
try:
async_llm = AsyncLLM.from_vllm_config(
vllm_config=vllm_config,
usage_context=usage_context,
enable_log_requests=engine_args.enable_log_requests,
disable_log_stats=engine_args.disable_log_stats,
client_addresses=client_config,
client_count=client_count,
client_index=client_index)
if disable_frontend_multiprocessing:
logger.warning(
"V1 is enabled, but got --disable-frontend-multiprocessing. "
"To disable frontend multiprocessing, set VLLM_USE_V1=0.")
# Don't keep the dummy data in memory
await async_llm.reset_mm_cache()
from vllm.v1.engine.async_llm import AsyncLLM
async_llm: Optional[AsyncLLM] = None
client_count = client_config.pop("client_count") if client_config else 1
client_index = client_config.pop("client_index") if client_config else 0
try:
async_llm = AsyncLLM.from_vllm_config(
vllm_config=vllm_config,
usage_context=usage_context,
enable_log_requests=engine_args.enable_log_requests,
disable_log_stats=engine_args.disable_log_stats,
client_addresses=client_config,
client_count=client_count,
client_index=client_index)
yield async_llm
finally:
if async_llm:
async_llm.shutdown()
# Don't keep the dummy data in memory
await async_llm.reset_mm_cache()
# V0 AsyncLLM.
else:
engine_client: Optional[EngineClient] = None
try:
engine_client = AsyncLLMEngine.from_vllm_config(
vllm_config=vllm_config,
usage_context=usage_context,
enable_log_requests=engine_args.enable_log_requests,
disable_log_stats=engine_args.disable_log_stats)
yield engine_client
finally:
if engine_client and hasattr(engine_client, "shutdown"):
engine_client.shutdown()
yield async_llm
finally:
if async_llm:
async_llm.shutdown()
async def validate_json_request(raw_request: Request):