A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 9, 18:15 UTC
[Bugfix][CI] Fix pre-commit (#32956)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Matthew Bonanni committed
Jan 23, 2026 at 13:26 UTC
2d6b537157af1ee2f65026b28a5258838e356f32
1 file changed
+1
-2
vllm/v1/worker/mamba_utils.py
+1
-2
@@ -4,13 +4,12 @@ import itertools
4
from typing import Any
5
6
import torch
7
-import triton
8
-import triton.language as tl
7
8
from vllm.config import CacheConfig
9
from vllm.model_executor.layers.mamba.mamba_utils import (
10
MambaStateCopyFunc,
11
)
12
+from vllm.triton_utils import tl, triton
13
from vllm.v1.core.sched.output import SchedulerOutput
14
from vllm.v1.kv_cache_interface import KVCacheConfig, MambaSpec
15
from vllm.v1.worker.gpu_input_batch import CachedRequestState