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

[V1] Refactor get_executor_cls (#11754)

Browse Source
This commit is contained in:
Rui Qiao
2025-01-05 23:59:16 -08:00
committed by GitHub
parent f8fcca100b
commit 022c5c6944
5 changed files with 26 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
vllm/v1/engine/async_llm.py
View File

@@ -22,7 +22,6 @@ from vllm.v1.engine.core_client import EngineCoreClient
from vllm.v1.engine.detokenizer import Detokenizer
from vllm.v1.engine.processor import Processor
from vllm.v1.executor.abstract import Executor
from vllm.v1.executor.ray_utils import initialize_ray_cluster
logger = init_logger(__name__)
@@ -105,7 +104,7 @@ class AsyncLLM(EngineClient):
else:
vllm_config = engine_config
executor_class = cls._get_executor_cls(vllm_config)
executor_class = Executor.get_class(vllm_config)
# Create the AsyncLLM.
return cls(
@@ -127,24 +126,6 @@ class AsyncLLM(EngineClient):
if handler := getattr(self, "output_handler", None):
handler.cancel()
@classmethod
def _get_executor_cls(cls, vllm_config: VllmConfig) -> Type[Executor]:
executor_class: Type[Executor]
distributed_executor_backend = (
vllm_config.parallel_config.distributed_executor_backend)
if distributed_executor_backend == "ray":
initialize_ray_cluster(vllm_config.parallel_config)
from vllm.v1.executor.ray_executor import RayExecutor
executor_class = RayExecutor
elif distributed_executor_backend == "mp":
from vllm.v1.executor.multiproc_executor import MultiprocExecutor
executor_class = MultiprocExecutor
else:
assert (distributed_executor_backend is None)
from vllm.v1.executor.uniproc_executor import UniprocExecutor
executor_class = UniprocExecutor
return executor_class
async def add_request(
self,
request_id: str,