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 13, 02:00 UTC
[CI/Build] Replace wikipedia url with local server ones (#28908)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Isotr0py committed
Nov 18, 2025 at 16:10 UTC
896e41ae04d18b0f984eefbb41b920aa7505f5d1
1 file changed
+11
-2
tests/entrypoints/openai/test_metrics.py
+11
-2
@@ -16,6 +16,7 @@ from transformers import AutoTokenizer
16
17
from vllm import version
18
19
+from ...conftest import LocalAssetServer
20
from ...utils import RemoteOpenAIServer
21
22
MODELS = {
@@ -69,7 +70,6 @@ async def client(server):
70
71
72
_PROMPT = "Hello my name is Robert and I love magic"
72
-_IMAGE_URL = "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
73
74
75
def _get_expected_values(num_requests: int, prompt_ids: list[int], max_tokens: int):
@@ -250,6 +250,7 @@ HIDDEN_DEPRECATED_METRICS: list[str] = [
250
251
@pytest.mark.asyncio
252
async def test_metrics_exist(
253
+ local_asset_server: LocalAssetServer,
254
server: RemoteOpenAIServer,
255
client: openai.AsyncClient,
256
model_key: str,
@@ -265,13 +266,21 @@ async def test_metrics_exist(
266
temperature=0.0,
267
)
268
else:
269
+ # https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg
270
await client.chat.completions.create(
271
model=model_name,
272
messages=[
273
{
274
"role": "user",
275
"content": [
274
- {"type": "image_url", "image_url": {"url": _IMAGE_URL}},
276
+ {
277
+ "type": "image_url",
278
+ "image_url": {
279
+ "url": local_asset_server.url_for(
280
+ "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
281
+ ),
282
+ },
283
+ },
284
{"type": "text", "text": "What's in this image?"},
285
],
286
}