[Bugfix] Qwen-vl output is inconsistent in speculative decoding

[skylee-01]

Qwen-vl uses speculative decoding and does not use speculative decoding to have inconsistent output. We tested ngram and medusa, and this situation exists. After investigation, we found that there was an error position_id, and further because the new seq_data was created during the batch_expansion process without passing mrope_position_delta information.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
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 do one of these:

• Add ready label to the PR
• Enable auto-merge.

🚀

[DarkLight1337]

Looks reasonable, thanks for fixing!