xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 06:45:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Misc] cache result of disable_inplace (#26666)

Browse Source 
Signed-off-by: Bill Nell <bnell@redhat.com>
This commit is contained in:
bnellnm 2025-10-12 20:17:50 -04:00 committed by GitHub
parent 7ef6052804
commit 60e419c1ee
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/layers/fused_moe/utils.py
View File

@ -1,5 +1,6 @@
# SPDX-License-Identifier: Apache-2.0 # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import functools
from math import prod from math import prod
import torch import torch
@ -325,5 +326,6 @@ def activation_without_mul(activation: str) -> str:
# Torch custom ops can't deal with outputs aliasing inputs so we need to # Torch custom ops can't deal with outputs aliasing inputs so we need to
# disable inplace for torch >= 2.9. # disable inplace for torch >= 2.9.
# See https://github.com/vllm-project/vllm/issues/26378 # See https://github.com/vllm-project/vllm/issues/26378
@functools.cache
def disable_inplace() -> bool: def disable_inplace() -> bool:
return is_torch_equal_or_newer("2.9") return is_torch_equal_or_newer("2.9")