[V0 Deprecation] Remove AsyncLLMEngine (#25025)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai> Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
@@ -7,7 +7,6 @@ import pytest
|
||||
import vllm.envs as envs
|
||||
from vllm import LLM
|
||||
from vllm.engine.arg_utils import AsyncEngineArgs
|
||||
from vllm.engine.async_llm_engine import AsyncLLMEngine
|
||||
|
||||
MODEL = "meta-llama/Llama-3.2-1B-Instruct"
|
||||
|
||||
@@ -96,20 +95,3 @@ def test_v1_attn_backend(monkeypatch):
|
||||
_ = AsyncEngineArgs(model=MODEL).create_engine_config()
|
||||
assert envs.VLLM_USE_V1
|
||||
m.delenv("VLLM_USE_V1")
|
||||
|
||||
|
||||
def test_reject_using_constructor_directly(monkeypatch):
|
||||
with monkeypatch.context() as m:
|
||||
if os.getenv("VLLM_USE_V1", None):
|
||||
m.delenv("VLLM_USE_V1")
|
||||
|
||||
# Sets VLLM_USE_V1=1.
|
||||
vllm_config = AsyncEngineArgs(model=MODEL).create_engine_config()
|
||||
|
||||
# This uses the V0 constructor directly.
|
||||
with pytest.raises(ValueError):
|
||||
AsyncLLMEngine(vllm_config,
|
||||
AsyncLLMEngine._get_executor_cls(vllm_config),
|
||||
log_stats=True)
|
||||
|
||||
m.delenv("VLLM_USE_V1")
|
||||
|
||||
Reference in New Issue
Block a user