[Klaud Cold] Update dsv4-fp4-b200-vllm (+mtp) vLLM image to v0.21.0#1476
[Klaud Cold] Update dsv4-fp4-b200-vllm (+mtp) vLLM image to v0.21.0#1476functionstackx wants to merge 2 commits into
Conversation
Update vLLM image from v0.20.0-cu130 (20d/18d old) to v0.21.0 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook 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 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. |
1 similar comment
|
Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook 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 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. |
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26010228368 |
There was a problem hiding this comment.
LGTM — straightforward vLLM image version bump for two dsv4-fp4-b200-vllm recipes with matching perf-changelog entry.
Extended reasoning...
Overview
This PR bumps the vLLM container image from vllm/vllm-openai:v0.20.0-cu130 to vllm/vllm-openai:v0.21.0 for two recipes (dsv4-fp4-b200-vllm and dsv4-fp4-b200-vllm-mtp) in .github/configs/nvidia-master.yaml, and adds the corresponding entry to perf-changelog.yaml. The total diff is 4 lines of meaningful change across 2 files.
Security risks
None. This is a config-only change touching image tags in YAML. No code, auth, crypto, or permission surfaces are involved. The image source (vllm/vllm-openai) is unchanged — only the tag is updated.
Level of scrutiny
Minimal scrutiny appropriate. This is a mechanical, well-established pattern (image version bump) that matches the repo's many prior image-bump PRs (e.g., the SGLang bump in #1458 mentioned in recent commits). The changelog entry is correctly formatted and the pr-link matches PR #1476.
Other factors
No bugs were flagged by the bug-hunting system. The full-sweep-enabled label indicates the change will be validated by a full perf sweep before merge. No outstanding reviewer comments — only the standard recipe-reminder bot notes.
Summary
Update vLLM image from v0.20.0-cu130 (20d/18d old) to v0.21.0
Recipes touched: `dsv4-fp4-b200-vllm`,
dsv4-fp4-b200-vllm-mtpTest plan
🤖 Generated with Claude Code