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 8, 16:30 UTC
[BugFix] Change support no act and mul for marlin (#34088)
Signed-off-by: Tomer Natan <tbarnatan@computelab-frontend-8.nvidia.com> Co-authored-by: Tomer Natan <tbarnatan@computelab-frontend-8.nvidia.com>
TomerBN-Nvidia committed
Feb 8, 2026 at 19:18 UTC
a263aa614060f8e6be52ed3de9995450d6c02892
1 file changed
+1
-1
vllm/model_executor/layers/fused_moe/fused_marlin_moe.py
+1
-1
@@ -561,7 +561,7 @@ class MarlinExpertsBase(mk.FusedMoEPermuteExpertsUnpermute):
561
562
@staticmethod
563
def _supports_no_act_and_mul() -> bool:
564
- return False
564
+ return True
565
566
@staticmethod
567
def _supports_quant_scheme(