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 10, 10:15 UTC
[fix] add cutedsl to global sf (#32001)
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
jiahanc committed
Jan 9, 2026 at 12:03 UTC
f9e2a75a1ee1d339ec5d885f842b3cfc27d71e02
1 file changed
+1
vllm/model_executor/layers/quantization/utils/flashinfer_utils.py
+1
@@ -264,6 +264,7 @@ def is_flashinfer_supporting_global_sf(backend: FlashinferMoeBackend | None) ->
264
backends_supporting_global_sf = (
265
FlashinferMoeBackend.CUTLASS,
266
FlashinferMoeBackend.TENSORRT_LLM,
267
+ FlashinferMoeBackend.CUTEDSL,
268
)
269
return backend in backends_supporting_global_sf
270