[Bugfix] Fixed the issue of not being able to input video and image simultaneously

[chaunceyjiang]

Fix #15291

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[ywang96]

Can confirm this PR fixes the issue - so I'm marking it as ready for 0.8.2 release.

[ywang96]

Client side code

import requests
import json

url = "http://127.0.0.1:8000/v1/chat/completions"
headers = {"Content-Type": "application/json", "Authorization": "Bearer token"}

model = "Qwen/Qwen2.5-VL-7B-Instruct"

image_url = "https://texascoffeeschool.com/wp-content/uploads/2021/10/DSC_0052-scaled.jpg"
video_url = "http://commondatastorage.googleapis.com/gtv-videos-bucket/sample/ForBiggerFun.mp4"

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "text",
                "text": "Describe all contents in a single sentence.",
            },
            {"type": "image_url", "image_url": {"url": image_url}},
            {"type": "video_url", "video_url": {"url": video_url}},
        ],
    },
]

data = {"model": model, "messages": messages}

response = requests.post(url, headers=headers, data=json.dumps(data))

Server-side log to verify the prompt is indeed parsed correctly

'<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\nDescribe the contents in a single sentence.<|vision_start|><|image_pad|><|vision_end|><|vision_start|><|video_pad|><|vision_end|><|im_end|>\n<|im_start|>assistant\n'

I don't think Qwen2.5VL is capable of mixed-modality inference, but at least this PR fixes the functionality to do so. Thanks for the work! @chaunceyjiang