xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-05 16:04:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge cc7d87c6dde84c9d86b247c575cd58a50711fed9 into 254f6b986720c92ddf97fbb1a6a6465da8e87e29

Browse Source
This commit is contained in:
Zhaodong Bing 2025-12-25 00:06:50 +00:00 committed by GitHub
commit 42e811f8bf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
View File

@ -211,16 +211,16 @@ class UnquantizedFusedMoEMethod(FusedMoEMethodBase, CustomOp):
super().process_weights_after_loading(layer) super().process_weights_after_loading(layer)
# Padding the weight for better performance on ROCm # Padding the weight for better performance on ROCm
layer.w13_weight.data = self._maybe_pad_weight(layer.w13_weight.data) layer.w13_weight.data.copy_(self._maybe_pad_weight(layer.w13_weight.data))
layer.w2_weight.data = self._maybe_pad_weight(layer.w2_weight.data) layer.w2_weight.data.copy_(self._maybe_pad_weight(layer.w2_weight.data))
if self.rocm_aiter_moe_enabled: if self.rocm_aiter_moe_enabled:
shuffled_w13, shuffled_w2 = rocm_aiter_ops.shuffle_weights( shuffled_w13, shuffled_w2 = rocm_aiter_ops.shuffle_weights(
layer.w13_weight.data, layer.w2_weight.data layer.w13_weight.data, layer.w2_weight.data
) )
layer.w13_weight.data = shuffled_w13 layer.w13_weight.data.copy_(shuffled_w13)
layer.w2_weight.data = shuffled_w2 layer.w2_weight.data.copy_(shuffled_w2)
if self.flashinfer_cutlass_moe_enabled: if self.flashinfer_cutlass_moe_enabled:
# Swap halves to arrange as [w3; w1] (kernel expectation) # Swap halves to arrange as [w3; w1] (kernel expectation)