[Attention] Update attention imports (#29540)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
This commit is contained in:
@@ -6,6 +6,7 @@ from typing import TYPE_CHECKING, Any, Optional, cast
|
||||
|
||||
import torch
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.linear import (
|
||||
@@ -102,8 +103,6 @@ class QuarkConfig(QuantizationConfig):
|
||||
def get_quant_method(
|
||||
self, layer: torch.nn.Module, prefix: str
|
||||
) -> Optional["QuantizeMethodBase"]:
|
||||
from vllm.attention.layer import Attention # Avoid circular import
|
||||
|
||||
# Check if the layer is skipped for quantization.
|
||||
exclude_layers = cast(list[str], self.quant_config.get("exclude"))
|
||||
if should_ignore_layer(
|
||||
|
||||
Reference in New Issue
Block a user