[Bugfix] Fix XliteModelRunner init failed when aclgraph is enabled (#5899)

### 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.

- vLLM version: v0.13.0
- vLLM main:
bde38c11df

Signed-off-by: changdawei1 <changdawei3@huawei.com>
This commit is contained in:
Magnus
2026-01-15 15:40:28 +08:00
committed by GitHub
parent 80fbb1b6b1
commit e8bbf72867

View File

@@ -2992,10 +2992,16 @@ class NPUModelRunner(GPUModelRunner):
set_draft_graph_params(self.cudagraph_batch_sizes)
def capture_model(self) -> None:
parent_module_name = self.__class__.__base__.__module__
gpu_model_runner_cls = next((cls for cls in self.__class__.__mro__
if cls.__name__ == "GPUModelRunner"),
None)
if gpu_model_runner_cls is None:
raise TypeError("Could not find GPUModelRunner in the MRO. "
"The class hierarchy may have changed.")
parent_module_name = gpu_model_runner_cls.__module__
with _torch_cuda_wrapper(), _replace_gpu_model_runner_function_wrapper(
parent_module_name):
super().capture_model()
GPUModelRunner.capture_model(self)
def _prepare_multimodal_fields(self):
"""