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

[CI] Add temperature=0.0, reduce max_tokens, and add debug prints to audio_in_video tests (#38492)

Browse Source 
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
This commit is contained in:
Andreas Karatzas
2026-03-30 00:36:45 -05:00
committed by GitHub
parent c133f33746
commit bea23536f6
1 changed files with 18 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
tests/entrypoints/openai/chat_completion/test_audio_in_video.py
View File

@@ -64,11 +64,12 @@ async def test_online_audio_in_video(
]
# multi-turn to test mm processor cache as well
for _ in range(2):
for turn in range(2):
chat_completion = await client.chat.completions.create(
model=MODEL_NAME,
messages=messages,
max_tokens=16,
max_tokens=8,
temperature=0.0,
extra_body={
"mm_processor_kwargs": {
"use_audio_in_video": True,
@@ -78,6 +79,12 @@ async def test_online_audio_in_video(
assert len(chat_completion.choices) == 1
choice = chat_completion.choices[0]
print(
f"[DEBUG][single-video] turn={turn} "
f"finish_reason={choice.finish_reason!r} "
f"content={choice.message.content!r} "
f"usage={chat_completion.usage}"
)
assert choice.finish_reason == "length"
@@ -111,11 +118,12 @@ async def test_online_audio_in_video_multi_videos(
]
# multi-turn to test mm processor cache as well
for _ in range(2):
for turn in range(2):
chat_completion = await client.chat.completions.create(
model=MODEL_NAME,
messages=messages,
max_tokens=16,
max_tokens=8,
temperature=0.0,
extra_body={
"mm_processor_kwargs": {
"use_audio_in_video": True,
@@ -125,6 +133,12 @@ async def test_online_audio_in_video_multi_videos(
assert len(chat_completion.choices) == 1
choice = chat_completion.choices[0]
print(
f"[DEBUG][multi-video] turn={turn} "
f"finish_reason={choice.finish_reason!r} "
f"content={choice.message.content!r} "
f"usage={chat_completion.usage}"
)
assert choice.finish_reason == "length"