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 4, 12:30 UTC
[CI] De-flake test_chat_completion_n_parameter_non_streaming (#41147)
Signed-off-by: haosdent <haosdent@gmail.com>
haosdent committed
Apr 29, 2026 at 11:23 UTC
75a7cf2c10f2dcc484c3e0444af33e0eaf3f4311
1 file changed
+3
-3
tests/entrypoints/openai/chat_completion/test_chat.py
+3
-3
@@ -845,9 +845,10 @@ async def test_chat_completion_n_parameter_non_streaming(
845
chat_completion = await client.chat.completions.create(
846
model=model_name,
847
messages=messages,
848
- max_completion_tokens=20,
849
- temperature=0.7,
848
+ max_completion_tokens=50,
849
+ temperature=1.0,
850
n=3,
851
+ seed=42,
852
stream=False,
853
)
854
@@ -859,7 +860,6 @@ async def test_chat_completion_n_parameter_non_streaming(
860
assert choice.message.content is not None
861
assert len(choice.message.content) > 0
862
862
- # Verify all responses are different (highly likely with temperature > 0)
863
contents = [choice.message.content for choice in chat_completion.choices]
864
assert len(set(contents)) > 1, "Expected different responses with n=3"
865