[CI] Block the cu126 wheel build while broken #23285
Conversation
Signed-off-by: mgoin <mgoin64@gmail.com>
|
👋 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 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 🚀 |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request aims to add a manual block for the CUDA 12.6 wheel build, which is a reasonable step if it's currently broken. However, I've found a critical issue in the implementation: a likely typo in a dependency key will prevent the block from working as intended. My review includes a specific code suggestion to correct this.
Signed-off-by: mgoin <mgoin64@gmail.com> Signed-off-by: Duncan Moss <djm.moss@gmail.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: mgoin <mgoin64@gmail.com> Signed-off-by: Xiao Yu <xiao.yu@amd.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
2 participants
Purpose
The CUDA 12.6 wheel build has been broken on the "release" job on main for several days due to a FlashAttention compilation issue on CUDA 12.6. We should disable it for now to keep main green.
https://buildkite.com/vllm/release/builds/7472/steps/canvas?sid=0198c912-0e89-4296-a799-9b8275153f55
Test Plan
Test Result
(Optional) Documentation Update
Essential Elements of an Effective PR Description Checklist
supported_models.mdandexamplesfor a new model.