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 15, 22:15 UTC
[Bugfix] Update expected token counts for Ultravox tests (#13895)
Cyrus Leung committed
Feb 26, 2025 at 20:56 UTC
934bb99c717434fe8533f1518a366d9f18196846
1 file changed
+3
-3
tests/entrypoints/openai/test_audio.py
+3
-3
@@ -83,7 +83,7 @@ async def test_single_chat_session_audio(client: openai.AsyncOpenAI,
83
choice = chat_completion.choices[0]
84
assert choice.finish_reason == "length"
85
assert chat_completion.usage == openai.types.CompletionUsage(
86
- completion_tokens=10, prompt_tokens=201, total_tokens=211)
86
+ completion_tokens=10, prompt_tokens=202, total_tokens=212)
87
88
message = choice.message
89
message = chat_completion.choices[0].message
@@ -140,7 +140,7 @@ async def test_single_chat_session_audio_base64encoded(
140
choice = chat_completion.choices[0]
141
assert choice.finish_reason == "length"
142
assert chat_completion.usage == openai.types.CompletionUsage(
143
- completion_tokens=10, prompt_tokens=201, total_tokens=211)
143
+ completion_tokens=10, prompt_tokens=202, total_tokens=212)
144
145
message = choice.message
146
message = chat_completion.choices[0].message
@@ -196,7 +196,7 @@ async def test_single_chat_session_input_audio(
196
choice = chat_completion.choices[0]
197
assert choice.finish_reason == "length"
198
assert chat_completion.usage == openai.types.CompletionUsage(
199
- completion_tokens=10, prompt_tokens=201, total_tokens=211)
199
+ completion_tokens=10, prompt_tokens=202, total_tokens=212)
200
201
message = choice.message
202
message = chat_completion.choices[0].message