[CI] Revert the C8 configuration for some test cases - #17505
kunpengW-code wants to merge 1 commit into
Conversation
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request updates several test configurations and documentation files by removing the --kv-cache-dtype int8 flag. Additionally, it updates the KV cache data type to fp8 for GLM5 models, ensuring consistency and alignment with current testing requirements. Highlights
New Features🧠 You can now enable Memory (public preview) to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize the Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counterproductive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
![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 updates the configuration for GLM models across documentation and end-to-end test files. Specifically, it removes the --kv-cache-dtype int8 and --attention_config.indexer_kv_dtype int8 parameters from GLM-5.1 and GLM-5.2 configurations, and switches the KV cache data type from int8 to fp8 in the GLM-5 documentation. The reviewer provided a structured PR title and summary suggestion that aligns with the repository's style guide.
github-actions
Bot
added
documentation
|
👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:
If CI fails, you can run linting and testing checks locally according Contributing and Testing. |
|
/nightly multi-node-GLM-5.1-W8A8C8-A3_128k_90_50
|
036b255 to
25fac7a
Compare
|
/nightly multi-node-GLM-5.1-W8A8C8-A3_128k_90_50
|
25fac7a to
0c4986f
Compare
|
/nightly multi-node-GLM-5.1-W8A8C8-A3_128k_90_50
|
0c4986f to
cb41c9d
Compare
|
/nightly multi-node-GLM-5.1-W8A8C8-A3_128k_90_50
|
cb41c9d to
facddab
Compare
|
/nightly multi-node-GLM-5.1-W8A8C8-A3_128k_90_50
|
Signed-off-by: wangkunpeng <1289706727@qq.com>
facddab to
64aeecc
Compare
|
/nightly multi-node-GLM-5.1-W8A8C8-A3_128k_90_50
|
|
This pull request has conflicts, please resolve those before we can evaluate the pull request. |
2 participants
What this PR does / why we need it?
In PR #15514, the startup configuration for C8 was updated. Some test cases still referenced the old configuration, so revert those test case changes.
This PR reverts the C8 configuration (specifically
--kv-cache-dtype int8and--attention_config.indexer_kv_dtype int8) for several GLM-5.1 and GLM-5.2 test cases. It also updates the documentation to usefp8instead ofint8for GLM-5/5.1 on 950DT Products, and removes theint8KV cache dtype configuration for GLM-5.2 where it is not recommended or has known issues.Does this PR introduce any user-facing change?
No, this is a test configuration and documentation update.
How was this patch tested?
Tested via existing E2E nightly and weekly multi-node CI pipelines.