[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
@@ -7,6 +7,7 @@ import torch
from torch.nn.parameter import Parameter
from vllm import _custom_ops as ops
from vllm.attention.layer import Attention
from vllm.logger import init_logger
from vllm.model_executor.layers.linear import LinearBase, UnquantizedLinearMethod
from vllm.model_executor.layers.quantization import QuantizationMethods
@@ -65,8 +66,6 @@ class PTPCFp8Config(Fp8Config):
def get_quant_method(
self, layer: torch.nn.Module, prefix: str
) -> Optional["QuantizeMethodBase"]:
from vllm.attention.layer import Attention # Avoid circular import
if isinstance(layer, LinearBase):
if is_layer_skipped(prefix, self.ignored_layers):
return UnquantizedLinearMethod()