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

Fix triton_kernels import error on some hardwares (#11831)

Browse Source
This commit is contained in:
fzyzcjy
2025-10-20 08:14:47 +08:00
committed by GitHub
parent 3b80232d06
commit a8ba32798e
1 changed files with 1 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
python/sglang/srt/layers/quantization/unquant.py
View File

@@ -1,6 +1,5 @@
from __future__ import annotations from __future__ import annotations
import importlib.util
from typing import TYPE_CHECKING, List, Optional from typing import TYPE_CHECKING, List, Optional
import torch import torch
@@ -31,8 +30,6 @@ if TYPE_CHECKING:
StandardDispatchOutput, StandardDispatchOutput,
) )
has_triton_kernels = importlib.util.find_spec("triton_kernels") is not None
_is_cpu_amx_available = cpu_has_amx_support() _is_cpu_amx_available = cpu_has_amx_support()
_is_hip = is_hip() _is_hip = is_hip()
@@ -143,7 +140,7 @@ class UnquantizedFusedMoEMethod(FusedMoEMethodBase, CustomOp):
self.triton_kernel_moe_forward = None self.triton_kernel_moe_forward = None
self.triton_kernel_moe_with_bias_forward = None self.triton_kernel_moe_with_bias_forward = None
if torch.cuda.is_available() and has_triton_kernels: if torch.cuda.is_available() and use_triton_kernels:
from sglang.srt.layers.moe.fused_moe_triton.triton_kernels_moe import ( from sglang.srt.layers.moe.fused_moe_triton.triton_kernels_moe import (
triton_kernel_moe_forward as _tk_forward, triton_kernel_moe_forward as _tk_forward,
) )