Merge branch 'vllm-project:main' into Eagle-mulitmodal-support-Qwen2.5vl

Author: LJH-LBJ

tests/entrypoints/openai/test_default_mm_loras.py

@@ -24,18 +24,7 @@
 
 
 @pytest.fixture(scope="module")
-def monkeypatch_module():
-    from _pytest.monkeypatch import MonkeyPatch
-    mpatch = MonkeyPatch()
-    yield mpatch
-    mpatch.undo()
-
-
-@pytest.fixture(scope="module", params=[False, True])
-def multimodal_server(request, monkeypatch_module):  # noqa: F811
-
-    use_v1 = request.param
-    monkeypatch_module.setenv('VLLM_USE_V1', '1' if use_v1 else '0')
+def multimodal_server():  # noqa: F811
 
     args = [
         # use half precision for speed and memory savings in CI environment