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 2, 11:00 UTC
[XPU] add gelu_tanh to xpu moe backend supported activations (#42822)
Signed-off-by: yintong-lu <yintong.lu@intel.com> Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
Yintong Lu committed
May 29, 2026 at 14:37 UTC
04516eabc840471c016c875991e1d72fb142c7c8
1 file changed
+1
vllm/model_executor/layers/fused_moe/experts/xpu_moe.py
+1
@@ -86,6 +86,7 @@ class XPUExperts(mk.FusedMoEExpertsModular):
86
return activation in [
87
MoEActivation.SILU,
88
MoEActivation.GELU,
89
+ MoEActivation.GELU_TANH,
90
MoEActivation.SWIGLUOAI,
91
MoEActivation.RELU2_NO_MUL,
92
]