xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-05 02:35:40 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Misc][gpt-oss] guard import when triton kernel when not up to date (#22584)

Browse Source 
Signed-off-by: zhewenli <zhewenli@meta.com>
This commit is contained in:
Zhewen Li 2025-08-10 21:29:35 -07:00 committed by GitHub
parent 1b99028069
commit afa5b7ca0b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 12 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
vllm/model_executor/layers/fused_moe/gpt_oss_triton_kernels_moe.py
View File

@ -5,15 +5,24 @@ from typing import TYPE_CHECKING, Any, Optional
import torch import torch
import vllm.model_executor.layers.fused_moe.modular_kernel as mk import vllm.model_executor.layers.fused_moe.modular_kernel as mk
from vllm.logger import init_logger
from vllm.model_executor.layers.fused_moe.topk_weight_and_reduce import ( from vllm.model_executor.layers.fused_moe.topk_weight_and_reduce import (
TopKWeightAndReduceDelegate) TopKWeightAndReduceDelegate)
from vllm.model_executor.layers.fused_moe.utils import extract_required_args from vllm.model_executor.layers.fused_moe.utils import extract_required_args
from vllm.utils import has_triton_kernels from vllm.utils import has_triton_kernels
logger = init_logger(__name__)
if has_triton_kernels(): if has_triton_kernels():
import triton_kernels.swiglu try:
from triton_kernels.matmul_ogs import FnSpecs, FusedActivation, matmul_ogs import triton_kernels.swiglu
from triton_kernels.routing import routing from triton_kernels.matmul_ogs import (FnSpecs, FusedActivation,
matmul_ogs)
from triton_kernels.routing import routing
except ModuleNotFoundError:
logger.error(
"Failed to import Triton kernels. Please make sure your triton "
"version is compatible.")
if TYPE_CHECKING: if TYPE_CHECKING:
from triton_kernels.matmul_ogs import PrecisionConfig from triton_kernels.matmul_ogs import PrecisionConfig