A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 04:30 UTC
[V1][Bugfix] Fix EngineCoreProc profile (#11185)
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Tyler Michael Smith committed
Dec 13, 2024 at 20:02 UTC
4b5b8a6a3bd94d9b0248b36b0eb4739d76fbb386
1 file changed
+1
-1
vllm/v1/engine/core.py
+1
-1
@@ -321,7 +321,7 @@ class EngineCoreProc(EngineCore):
321
if isinstance(request, EngineCoreRequest):
322
self.add_request(request)
323
elif isinstance(request, EngineCoreProfile):
324
- self.model_executor.worker.profile(request.is_start)
324
+ self.model_executor.profile(request.is_start)
325
else:
326
# TODO: make an EngineCoreAbort wrapper
327
assert isinstance(request, list)