Upgrade vllm commit hash to 1216 #5053
Conversation
![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 vLLM commit hash in the versioning_policy.md documentation file. I have added one suggestion to format the commit hash as a clickable link to improve the documentation's usability and help prevent potential issues for users. Additionally, for future pull requests, please provide a descriptive title and fill out the description template to give reviewers better context, as the current title is unclear and the description is empty.
| | vLLM Ascend | vLLM | Python | Stable CANN | PyTorch/torch_npu | | ||
| |-------------|--------------|------------------|-------------|--------------------| | ||
| | main | 4429d934de3c5cc327b0d7aec8e473aeba38db90, v0.12.0 tag | >= 3.10, < 3.12 | 8.3.RC2 | 2.8.0 / 2.8.0 | | ||
| | main | 6063853ead105f89400770b4fe6760959796d827, v0.12.0 tag | >= 3.10, < 3.12 | 8.3.RC2 | 2.8.0 / 2.8.0 | |
There was a problem hiding this comment.
Hard-coding a full 40-character commit hash is error-prone. A typo could lead users to an incorrect commit, causing significant issues when they try to set up their environment based on this documentation. To improve usability and reduce this risk, I suggest making this hash a clickable link to the commit on GitHub. This allows for easy verification.
| | main | 6063853ead105f89400770b4fe6760959796d827, v0.12.0 tag | >= 3.10, < 3.12 | 8.3.RC2 | 2.8.0 / 2.8.0 | | |
| | main | [6063853e](https://github.com/vllm-project/vllm/commit/6063853ead105f89400770b4fe6760959796d827), v0.12.0 tag | >= 3.10, < 3.12 | 8.3.RC2 | 2.8.0 / 2.8.0 | |
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. |
vllm_ascend/platform.py
Outdated
| use_sparse = kwargs["attn_selector_config"].use_sparse | ||
| else: | ||
| use_mla = kwargs.get("use_mla", args[4] if len(args) >= 5 else None) | ||
| use_sparse = kwargs.get("use_sparse", args[5] if len(args) >= 6 else None) |
Please paste the PR in vllm than break vllm-ascend |
wangxiyuan
added
ready
|
please fix the CI error |
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: zxwang <[email protected]>
Signed-off-by: leo-pony <[email protected]>
Signed-off-by: leo-pony <[email protected]>
Signed-off-by: leo-pony <[email protected]>
Signed-off-by: leo-pony <[email protected]>
Signed-off-by: leo-pony <[email protected]>
|
I‘ll fix the UT later |
Fix broken ut introduced by #5053 - vLLM version: v0.12.0 - vLLM main: vllm-project/vllm@ad32e3e Signed-off-by: wangxiyuan <[email protected]>
3 participants
What this PR does / why we need it?
Upstream vLLM PR #30212 vllm-project/vllm#30212 and vllm-project/vllm#29646
refactored the attention backend selection interface, This PR adapts vllm-ascend's get_attn_backend_cls to align with the new upstream standard, ensuring compatibility and reducing maintenance overhead.
Does this PR introduce any user-facing change?
How was this patch tested?
co-author:[leo-pony][email protected]