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 11, 10:00 UTC
update piecewise cudagraph warning when splitting_ops=[] (#30728)
Signed-off-by: Boyuan Feng <boyuan@meta.com>
Boyuan Feng committed
Dec 16, 2025 at 06:09 UTC
104003dc77d7d532ea6f946a6ea72ef8ea749078
1 file changed
+1
-1
vllm/config/compilation.py
+1
-1
@@ -937,7 +937,7 @@ class CompilationConfig:
937
or self.cudagraph_mode == CUDAGraphMode.FULL_AND_PIECEWISE
938
):
939
logger.warning_once(
940
- "Using piecewise compilation with empty splitting_ops"
940
+ "Using piecewise cudagraph with empty splitting_ops"
941
)
942
if self.cudagraph_mode == CUDAGraphMode.PIECEWISE:
943
logger.warning_once(