docs(vllm-omni): note ARM64 not supported; flag GLM-Image as experimental

[ayushag-nv]

Overview:

Doc-only update to docs/backends/vllm/vllm-omni.md:

• Note that vLLM-Omni is not currently installed on arm64 builds.
• Flag zai-org/GLM-Image as experimental — generation may fail or produce incorrect outputs for some prompts/sizes.

No code changes.

Details:

• Under ## Prerequisites → ### Installation, add an "ARM64 not supported" callout: the container build skips the vLLM-Omni install on arm64, so vLLM-Omni features are unavailable on that arch today.
• Under ### Quick Start: GLM-Image (2-Stage, 2 GPUs), add an "Experimental" callout above the existing transformers>=5.0 known-issue note, warning that generation may fail or produce incorrect/garbled outputs for some prompts and sizes.

Where should the reviewer start?

• docs/backends/vllm/vllm-omni.md — the only file changed. +4 lines.

Related Issues: (use one of the action keywords Closes / Fixes / Resolves / Relates to)

• Relates to DYN-2741: dynamo/release-1.1.0/vllm-runtime/arm64 — vllm-omni module missing from arm64

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 1 additional finding.

[ayushag-nv]

/ok to test 6e3eb78

[coderabbitai]

Walkthrough

The vLLM-Omni installation gating logic was broadened to trigger whenever VLLM_OMNI_REF is non-empty, removing the prior requirement for ARCH=amd64. Status messages were updated to include architecture information, and the skip message was revised.

Changes

Cohort / File(s)	Summary
vLLM Installation Gating container/deps/vllm/install_vllm.sh	Relaxed installation condition from requiring both VLLM_OMNI_REF and ARCH=amd64 to only VLLM_OMNI_REF non-empty. Updated status messages to include arch=${ARCH} for PyPI success and source build fallback. Removed ARM64-not-supported skip message.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 1 | ❌ 2

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Title check	⚠️ Warning	The PR title mentions 'docs' and ARM64/GLM-Image documentation changes, but the raw summary shows code changes to install_vllm.sh script, not documentation-only changes.	Update the title to accurately reflect the actual code changes (e.g., 'chore(container): broaden vLLM-Omni installation gating logic') or clarify if only docs were modified.
Description check	⚠️ Warning	The PR description is incomplete and contradictory. It claims 'No code changes' but the raw summary shows 7 lines added/6 removed in vllm install script.	Update the description to accurately reflect the actual code changes in container/deps/vllm/install_vllm.sh and clarify whether the PR modifies documentation, code, or both.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[ayushag-nv]

/ok to test 9005a8d

[github-actions]

🌿 Fern Docs Preview: https://nvidia-preview-eabe7e44-25cb-4394-a43e-5e191d156519.docs.buildwithfern.com/dynamo/dev

[rmccorm4]

Two small notes

[ayushag-nv]

/ok to test e5f8f45