Update kimik2.5-int4-mi300x-vllm vLLM image to v0.20.2

[Klaud-Cold]

Summary

• Update kimik2.5-int4-mi300x-vllm image from vllm/vllm-openai-rocm:v0.18.0 to vllm/vllm-openai-rocm:v0.20.2

Ref #1154

Generated with Claude Code

[github-actions]

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

• https://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

• https://github.com/vllm-project/recipes
• https://github.com/sgl-project/sglang/tree/main/docs_new

PR authors are responsible for ensuring that after merging, all GitHub Action jobs fully pass. A lot of the time, failures are just flakes and simply re-running the failed jobs will fix it. If re-running failed jobs is attempted, PR authors are responsible for ensuring it passes. See GitHub's docs on re-running failed jobs: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/re-run-workflows-and-jobs#re-running-failed-jobs-in-a-workflow

As a rule of thumb, generally, PR authors should request a review & get a PR approval from the respective companies' CODEOWNERS before requesting a review from core maintainers.

If additional help is needed, PR authors can reach out to core maintainers over Slack.

[github-actions]

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

• https://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

• https://github.com/vllm-project/recipes
• https://github.com/sgl-project/sglang/tree/main/docs_new

PR authors are responsible for ensuring that after merging, all GitHub Action jobs fully pass. A lot of the time, failures are just flakes and simply re-running the failed jobs will fix it. If re-running failed jobs is attempted, PR authors are responsible for ensuring it passes. See GitHub's docs on re-running failed jobs: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/re-run-workflows-and-jobs#re-running-failed-jobs-in-a-workflow

As a rule of thumb, generally, PR authors should request a review & get a PR approval from the respective companies' CODEOWNERS before requesting a review from core maintainers.

If additional help is needed, PR authors can reach out to core maintainers over Slack.

[claude]

🟡 The new perf-changelog entry at line 2351 has pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX — the literal XXX placeholder from the PR template was not replaced with the actual PR number. This will produce a 404 link in the changelog; fix by replacing XXX with 1350.

Extended reasoning...

What's wrong

perf-changelog.yaml line 2351 reads:

pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX

The XXX is a stub from the "Updating Docker Images" section of AGENTS.md, where the template literally shows pull/XXX as a placeholder for authors to fill in. The author/generator of this PR copied the template but didn't substitute the actual PR number.

Why this is the only such case

Every other entry in perf-changelog.yaml uses a real numeric PR id — e.g. line 2332 → pull/1305, line 2338 → pull/1308, line 2345 → pull/1310. A grep for non-numeric pr-link values turns up exactly one hit: the new entry added by this PR. So this is clearly an authoring oversight, not a convention.

Impact

When this is merged, the changelog entry for the kimik2.5-int4-mi300x-vllm v0.18.0 → v0.20.2 image bump will link to github.com/SemiAnalysisAI/InferenceX/pull/XXX, which 404s. Anyone using the changelog to trace a performance change back to its PR for this entry will hit a dead link, breaking changelog→PR traceability. No runtime/benchmark impact — purely a metadata/documentation issue, hence nit severity.

Step-by-step proof

1. This PR is Update kimik2.5-int4-mi300x-vllm vLLM image to v0.20.2 #1350 (per PR metadata: <pr number="1350">).
2. The diff adds a new perf-changelog.yaml entry whose pr-link ends in /pull/XXX (diff hunk at perf-changelog.yaml:2346–2351).
3. XXX is not a valid GitHub PR number; navigating to https://github.com/SemiAnalysisAI/InferenceX/pull/XXX returns a 404.
4. All other recent entries (lines 2332, 2338, 2345) use real numeric ids, confirming the expected format.
5. Replacing XXX with 1350 (this PR's number) yields a working link to the merge commit, restoring traceability.

Fix

-  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1350

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25763476065
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25763476065

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25763476065
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25763476065