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, 00:45 UTC
[chore] Move the rest of wikimedia url to S3 (#28921)
Signed-off-by: Kevin H. Luu <khluu000@gmail.com> Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Kevin H. Luu committed
Nov 18, 2025 at 09:44 UTC
c64c0b78de4716ef019666663c56b6ceaa019463
11 files changed
+24
-24
docs/features/multimodal_inputs.md
+1
-1
@@ -483,7 +483,7 @@ Then, you can use the OpenAI client as follows:
483
)
484
485
# Single-image input inference
486
- 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"
486
+ image_url = "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
487
488
chat_response = client.chat.completions.create(
489
model="microsoft/Phi-3.5-vision-instruct",
docs/serving/openai_compatible_server.md
+1
-1
@@ -293,7 +293,7 @@ and passing a list of `messages` in the request. Refer to the examples below for
293
base_url="http://localhost:8000/v1",
294
api_key="EMPTY",
295
)
296
- 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"
296
+ image_url = "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
297
298
response = create_chat_embeddings(
299
client,
examples/offline_inference/vision_language_pooling.py
+2
-2
@@ -266,7 +266,7 @@ def get_query(modality: QueryModality):
266
return ImageQuery(
267
modality="image",
268
image=fetch_image(
269
- "https://upload.wikimedia.org/wikipedia/commons/thumb/4/47/American_Eskimo_Dog.jpg/360px-American_Eskimo_Dog.jpg" # noqa: E501
269
+ "https://vllm-public-assets.s3.us-west-2.amazonaws.com/multimodal_asset/eskimo.jpg" # noqa: E501
270
),
271
)
272
@@ -275,7 +275,7 @@ def get_query(modality: QueryModality):
275
modality="text+image",
276
text="A cat standing in the snow.",
277
image=fetch_image(
278
- "https://upload.wikimedia.org/wikipedia/commons/thumb/b/b6/Felis_catus-cat_on_snow.jpg/179px-Felis_catus-cat_on_snow.jpg" # noqa: E501
278
+ "https://vllm-public-assets.s3.us-west-2.amazonaws.com/multimodal_asset/cat_snow.jpg" # noqa: E501
279
),
280
)
281
examples/online_serving/openai_chat_completion_client_for_multimodal.py
+1
-1
@@ -66,7 +66,7 @@ def run_text_only(model: str, max_completion_tokens: int) -> None:
66
# Single-image input inference
67
def run_single_image(model: str, max_completion_tokens: int) -> None:
68
## Use image url in the payload
69
- 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"
69
+ image_url = "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
70
chat_completion_from_url = client.chat.completions.create(
71
messages=[
72
{
examples/online_serving/pooling/openai_chat_embedding_client_for_multimodal.py
+1
-1
@@ -21,7 +21,7 @@ from PIL import Image
21
openai_api_key = "EMPTY"
22
openai_api_base = "http://localhost:8000/v1"
23
24
-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"
24
+image_url = "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
25
26
27
def create_chat_embeddings(
tests/entrypoints/openai/test_vision.py
+4
-4
@@ -17,10 +17,10 @@ MAXIMUM_IMAGES = 2
17
18
# Test different image extensions (JPG/PNG) and formats (gray/RGB/RGBA)
19
TEST_IMAGE_ASSETS = [
20
- "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
21
- "Grayscale_8bits_palette_sample_image.png", # "https://upload.wikimedia.org/wikipedia/commons/f/fa/Grayscale_8bits_palette_sample_image.png",
22
- "1280px-Venn_diagram_rgb.svg.png", # "https://upload.wikimedia.org/wikipedia/commons/thumb/9/91/Venn_diagram_rgb.svg/1280px-Venn_diagram_rgb.svg.png",
23
- "RGBA_comp.png", # "https://upload.wikimedia.org/wikipedia/commons/0/0b/RGBA_comp.png",
20
+ "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
21
+ "Grayscale_8bits_palette_sample_image.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/Grayscale_8bits_palette_sample_image.png",
22
+ "1280px-Venn_diagram_rgb.svg.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/1280px-Venn_diagram_rgb.svg.png",
23
+ "RGBA_comp.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/RGBA_comp.png",
24
]
25
26
EXPECTED_MM_BEAM_SEARCH_RES = [
tests/entrypoints/pooling/openai/test_vision_embedding.py
+4
-4
@@ -19,10 +19,10 @@ assert vlm2vec_jinja_path.exists()
19
20
# Test different image extensions (JPG/PNG) and formats (gray/RGB/RGBA)
21
TEST_IMAGE_ASSETS = [
22
- "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
23
- "Grayscale_8bits_palette_sample_image.png", # "https://upload.wikimedia.org/wikipedia/commons/f/fa/Grayscale_8bits_palette_sample_image.png",
24
- "1280px-Venn_diagram_rgb.svg.png", # "https://upload.wikimedia.org/wikipedia/commons/thumb/9/91/Venn_diagram_rgb.svg/1280px-Venn_diagram_rgb.svg.png",
25
- "RGBA_comp.png", # "https://upload.wikimedia.org/wikipedia/commons/0/0b/RGBA_comp.png",
22
+ "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
23
+ "Grayscale_8bits_palette_sample_image.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/Grayscale_8bits_palette_sample_image.png",
24
+ "1280px-Venn_diagram_rgb.svg.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/1280px-Venn_diagram_rgb.svg.png",
25
+ "RGBA_comp.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/RGBA_comp.png",
26
]
27
28
tests/models/language/pooling/test_mm_classifier_conversion.py
+1
-1
@@ -75,7 +75,7 @@ def test_gemma_multimodal(
75
{
76
"type": "image_url",
77
"image_url": {
78
- "url": "https://upload.wikimedia.org/wikipedia/commons/c/c6/Set_of_fourteen_side_chairs_MET_DP110780.jpg"
78
+ "url": "https://vllm-public-assets.s3.us-west-2.amazonaws.com/multimodal_asset/red_chair.jpg"
79
},
80
},
81
{"type": "text", "text": "A fine 19th century piece of furniture."},
tests/multimodal/test_utils.py
+4
-4
@@ -16,10 +16,10 @@ from vllm.multimodal.utils import MediaConnector, argsort_mm_positions
16
17
# Test different image extensions (JPG/PNG) and formats (gray/RGB/RGBA)
18
TEST_IMAGE_ASSETS = [
19
- "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
20
- "Grayscale_8bits_palette_sample_image.png", # "https://upload.wikimedia.org/wikipedia/commons/f/fa/Grayscale_8bits_palette_sample_image.png",
21
- "1280px-Venn_diagram_rgb.svg.png", # "https://upload.wikimedia.org/wikipedia/commons/thumb/9/91/Venn_diagram_rgb.svg/1280px-Venn_diagram_rgb.svg.png",
22
- "RGBA_comp.png", # "https://upload.wikimedia.org/wikipedia/commons/0/0b/RGBA_comp.png",
19
+ "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
20
+ "Grayscale_8bits_palette_sample_image.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/Grayscale_8bits_palette_sample_image.png",
21
+ "1280px-Venn_diagram_rgb.svg.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/1280px-Venn_diagram_rgb.svg.png",
22
+ "RGBA_comp.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/RGBA_comp.png",
23
]
24
25
TEST_VIDEO_URLS = [
tests/utils.py
+1
-1
@@ -676,7 +676,7 @@ def compare_all_settings(
676
results += _test_image_text(
677
client,
678
model,
679
- "https://upload.wikimedia.org/wikipedia/commons/0/0b/RGBA_comp.png",
679
+ "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/RGBA_comp.png",
680
)
681
elif method == "encode":
682
results += _test_embeddings(client, model, prompt)
tests/v1/entrypoints/openai/serving_responses/test_image.py
+4
-4
@@ -15,10 +15,10 @@ MODEL_NAME = "Qwen/Qwen2.5-VL-3B-Instruct"
15
MAXIMUM_IMAGES = 2
16
# Test different image extensions (JPG/PNG) and formats (gray/RGB/RGBA)
17
TEST_IMAGE_ASSETS = [
18
- "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
19
- "Grayscale_8bits_palette_sample_image.png", # "https://upload.wikimedia.org/wikipedia/commons/f/fa/Grayscale_8bits_palette_sample_image.png",
20
- "1280px-Venn_diagram_rgb.svg.png", # "https://upload.wikimedia.org/wikipedia/commons/thumb/9/91/Venn_diagram_rgb.svg/1280px-Venn_diagram_rgb.svg.png",
21
- "RGBA_comp.png", # "https://upload.wikimedia.org/wikipedia/commons/0/0b/RGBA_comp.png",
18
+ "2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
19
+ "Grayscale_8bits_palette_sample_image.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/Grayscale_8bits_palette_sample_image.png",
20
+ "1280px-Venn_diagram_rgb.svg.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/1280px-Venn_diagram_rgb.svg.png",
21
+ "RGBA_comp.png", # "https://vllm-public-assets.s3.us-west-2.amazonaws.com/vision_model_images/RGBA_comp.png",
22
]
23
24