[v0.13.0][Bugfix] Fix XliteModelRunner init failed when aclgraph is enabled#5887
Merged
yiz-liu merged 1 commit intovllm-project:releases/v0.13.0from Jan 21, 2026
Merged
Conversation
changdawei1
changed the title
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request addresses a bug in XliteModelRunner that led to an initialization failure when aclgraph is enabled. The problem was that the inherited capture_model method was incorrectly identifying a module for patching, which is essential for Ascend NPU graph capturing. The fix correctly overrides capture_model in XliteModelRunner to locate the appropriate module. The change is correct and necessary. I have included one suggestion to enhance the implementation's robustness and maintainability by avoiding string-based class lookups.
Comment on lines
+44
to
+49
| gpu_runner_cls = next((cls for cls in self.__class__.__mro__ | ||
| if cls.__name__ == "GPUModelRunner"), None) | ||
| if gpu_runner_cls is None: | ||
| raise TypeError("Could not find GPUModelRunner in the MRO. " | ||
| "The class hierarchy may have changed.") | ||
| parent_module_name = gpu_runner_cls.__module__ |
Contributor
There was a problem hiding this comment.
Since GPUModelRunner is imported, you can simplify this logic by using the class object directly instead of searching for it in the MRO by its string name. This approach is more robust and less likely to break if the class is ever renamed.
You can replace this block with a simple isinstance check and direct access to GPUModelRunner.__module__.
if not isinstance(self, GPUModelRunner):
raise TypeError("Could not find GPUModelRunner in the MRO. "
"The class hierarchy may have changed.")
parent_module_name = GPUModelRunner.__module__
changdawei1
force-pushed
the
v0.13.0
branch
3 times, most recently
from
c848e46 to
891fc7c
Compare
Signed-off-by: changdawei1 <changdawei3@huawei.com>
845473182
pushed a commit
to 845473182/vllm-ascend
that referenced
this pull request
Jan 21, 2026
…lm-ascend into FIA_v0.13.0 * 'releases/v0.13.0' of https://github.com/vllm-project/vllm-ascend: [0.13.0][Bugfix] Fix setting of `speculative_config.enforce_eager` for dsv32 (vllm-project#5958) [v0.13.0][Bugfix] Fix XliteModelRunner init failed when aclgraph is enabled (vllm-project#5887) [0.13.0][Bugfix] Fixed an problem related to embeddings sharing (vllm-project#5972) [Bugfix]Fixed precision issues caused by pooled request pooling (vllm-project#6057) [0.13.0][Bugfix] fix pcp aclgraph qwen FIA bug (vllm-project#6038) [0.13.0][cherry-pick][bugfix] fix bug of triton mrope (vllm-project#6009) 【0.13.0】【bugfix】Resolved memory deallocation failure in the pooling layer under re-computation workloads. (vllm-project#6056)
tangtiangu
pushed a commit
to tangtiangu/jiusi-vllm-ascend
that referenced
this pull request
Feb 24, 2026
…nabled (vllm-project#5887) ### What this PR does / why we need it? Fix XliteModelRunner init failed when aclgraph is enabled. Ensure function graph_capture of vllm.v1.worker.gpu_model_runner is replaced. ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? Signed-off-by: changdawei1 <changdawei3@huawei.com>
tangtiangu
pushed a commit
to tangtiangu/jiusi-vllm-ascend
that referenced
this pull request
Feb 24, 2026
…nabled (vllm-project#5887) ### What this PR does / why we need it? Fix XliteModelRunner init failed when aclgraph is enabled. Ensure function graph_capture of vllm.v1.worker.gpu_model_runner is replaced. ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? Signed-off-by: changdawei1 <changdawei3@huawei.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this PR does / why we need it?
Fix XliteModelRunner init failed when aclgraph is enabled. Ensure function graph_capture of vllm.v1.worker.gpu_model_runner is replaced.
Does this PR introduce any user-facing change?
How was this patch tested?