minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

Fix the accuracy issue when running mxfp4 dsv3 model and enable ep (#15304)

Browse Source 
Co-authored-by: wunhuang <wunhuang@amd.com>
This commit is contained in:
kk
2025-12-17 01:53:22 -08:00
committed by GitHub
co-authored by wunhuang
parent e9abb52576
commit 8c34e18140
2 changed files with 2 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/layers/quantization/mxfp4.py
+1
View File
@@ -866,5 +866,6 @@ class Mxfp4DynamicQuantMoEMethod(FusedMoEMethodBase):
else ActivationType.Gelu else ActivationType.Gelu
), ),
doweight_stage1=False, doweight_stage1=False,
expert_mask=layer.expert_mask_gpu,
) )
return StandardCombineInput(hidden_states=output) return StandardCombineInput(hidden_states=output)
python/sglang/srt/layers/quantization/quark/quark_moe.py
+1
View File
@@ -245,6 +245,7 @@ class QuarkW4A4MXFp4MoEMethod(QuarkMoEMethod):
else ActivationType.Gelu else ActivationType.Gelu
), ),
doweight_stage1=False, doweight_stage1=False,
expert_mask=layer.expert_mask_gpu,
) )
return StandardCombineInput(hidden_states=output) return StandardCombineInput(hidden_states=output)