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 15, 18:30 UTC
[V1] Fix wrong import path of get_flash_attn_version (#15280)
Signed-off-by: Lehua Ding <lehuading@tencent.com>
Lehua Ding committed
Mar 21, 2025 at 18:54 UTC
91ca929dc7aa626c926cd992b62a45b47e610b6a
1 file changed
+1
-1
vllm/v1/attention/backends/mla/common.py
+1
-1
@@ -195,8 +195,8 @@ from vllm import _custom_ops as ops
195
from vllm.attention.backends.abstract import (AttentionBackend, AttentionLayer,
196
AttentionMetadata,
197
MLAAttentionImpl)
198
-from vllm.attention.backends.utils import get_flash_attn_version
198
from vllm.attention.ops.triton_merge_attn_states import merge_attn_states
199
+from vllm.fa_utils import get_flash_attn_version
200
from vllm.logger import init_logger
201
from vllm.model_executor.layers.linear import (ColumnParallelLinear,
202
LinearBase, RowParallelLinear,