EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Fix global input scale incompatible with CuTe DSL moe (#10370)

Browse Source
This commit is contained in:
fzyzcjy
2025-09-12 18:22:49 +08:00
committed by GitHub
parent 3a77c80b26
commit efedbe6ca9
1 changed files with 15 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
python/sglang/srt/layers/quantization/modelopt_quant.py
View File

@@ -1187,6 +1187,21 @@ class ModelOptNvFp4FusedMoEMethod(FusedMoEMethodBase):
if self.enable_flashinfer_cutlass_moe or self.enable_flashinfer_trtllm_moe:
w13_input_scale = layer.w13_input_scale.max().to(torch.float32)
w2_input_scale = layer.w2_input_scale.max().to(torch.float32)
elif self.enable_flashinfer_cutedsl_moe:
w13_input_scale = layer.w13_input_scale.max(dim=1).values.to(torch.float32)
w2_input_scale = layer.w2_input_scale
def _slice_scale(w):
assert w.shape == (layer.num_experts,)
assert layer.moe_ep_size * layer.num_local_experts == layer.num_experts
return w[
layer.moe_ep_rank
* layer.num_local_experts : (layer.moe_ep_rank + 1)
* layer.num_local_experts
]
w13_input_scale = _slice_scale(w13_input_scale)
w2_input_scale = _slice_scale(w2_input_scale)
else:
w13_input_scale = layer.w13_input_scale.max(dim=1).values.to(torch.float32)
w2_input_scale = layer.w2_input_scale