update benchmark_serving_structured_output to include auto backend

[Chenyaaang]

follow up #15724

After setting structured output backend to auto, update the benchmark script to also accept "auto" as structured-output-backend.

[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.

🚀

[mgoin]

Oh I see now, nice! Auto is now the default for both v0 and v1 right?

[russellb]

Oh I see now, nice! Auto is now the default for both v0 and v1 right?

Yeah. In V0, auto is an alias for xgrammar. In V1, auto is the default and xgrammar is with no fallback.

[Chenyaaang]

Oh I see now, nice! Auto is now the default for both v0 and v1 right?

Yes, my understanding is v0 and v1 can both use auto. In v0 auto will fallback to xgrammar, while in V1 auto will select from xgrammar and guidance.