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 7, 04:00 UTC
[Bugfix] Fix simple Mistral-Small example (#36156)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Mar 6, 2026 at 12:25 UTC
de00ebeac4abddafff9f23bb598a6619b5892261
1 file changed
+5
-2
examples/offline_inference/mistral-small.py
+5
-2
@@ -7,6 +7,7 @@ import argparse
7
from vllm import LLM
8
from vllm.sampling_params import SamplingParams
9
from vllm.assets.image import ImageAsset
10
+from vllm.multimodal.utils import encode_image_url
11
12
# This script is an offline demo for running Mistral-Small-3.1
13
#
@@ -79,8 +80,10 @@ def run_simple_demo(args: argparse.Namespace):
80
"content": [
81
{"type": "text", "text": prompt},
82
{
82
- "type": "image_pil",
83
- "image_pil": ImageAsset("cherry_blossom").pil_image,
83
+ "type": "image_url",
84
+ "image_url": {
85
+ "url": encode_image_url(ImageAsset("cherry_blossom").pil_image)
86
+ },
87
},
88
],
89
},