[Attention] Update attention imports (#29540)

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
This commit is contained in:
Matthew Bonanni
2025-11-27 11:19:09 -05:00
committed by GitHub
parent cd007a53b4
commit fc1d8be3dc
38 changed files with 63 additions and 126 deletions
+1 -4
View File
@@ -12,12 +12,12 @@ from typing import TYPE_CHECKING, Any, NamedTuple
import numpy as np
import torch
from vllm.attention.backends.registry import AttentionBackendEnum
from vllm.logger import init_logger
if TYPE_CHECKING:
from torch.distributed import PrefixStore, ProcessGroup
from vllm.attention.backends.registry import AttentionBackendEnum
from vllm.config import VllmConfig
from vllm.config.cache import CacheDType
from vllm.inputs import ProcessorInputs, PromptType
@@ -226,9 +226,6 @@ class Platform:
def get_vit_attn_backend(
cls, head_size: int, dtype: torch.dtype
) -> "AttentionBackendEnum":
# Import AttentionBackendEnum here to avoid circular import.
from vllm.attention.backends.registry import AttentionBackendEnum
return AttentionBackendEnum.TORCH_SDPA
@classmethod