A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 2, 07:15 UTC
[XPU][CI] Fix test_audio_in_video flake by using module-scoped server fixture (#44146)
Signed-off-by: Chaojun Zhang <chaojun.zhang@intel.com>
Chaojun Zhang committed
Jun 1, 2026 at 19:21 UTC
bd0aecdc087382c2e3411c24e5d252d8b83cfe25
1 file changed
+5
-1
tests/entrypoints/openai/chat_completion/test_audio_in_video.py
+5
-1
@@ -14,8 +14,12 @@ from tests.utils import ROCM_EXTRA_ARGS, RemoteOpenAIServer
14
MODEL_NAME = "Qwen/Qwen2.5-Omni-3B"
15
16
17
-@pytest.fixture
17
+@pytest.fixture(scope="module")
18
def server():
19
+ # Use module scope so the server is started once and shared across all
20
+ # tests in this file. Starting a new vLLM server per test on XPU can
21
+ # cause the second server startup to hang silently and exceed the
22
+ # wait-for-server timeout, resulting in RuntimeError.
23
args = [
24
"--max-model-len",
25
"16384",