[Attention] Refactor CUDA attention backend selection logic (#24794)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com> Signed-off-by: Matthew Bonanni <mbonanni001@gmail.com> Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
This commit is contained in:
co-authored by
Luka Govedič
parent
2e78150d24
commit
b30dfa03c5
@@ -10,7 +10,7 @@ import torch
|
||||
import torch.nn as nn
|
||||
from transformers import BaseImageProcessor, BatchFeature, PretrainedConfig
|
||||
|
||||
from vllm.attention.backends.registry import _Backend
|
||||
from vllm.attention.backends.registry import AttentionBackendEnum
|
||||
from vllm.config import VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.model_executor.layers.linear import ReplicatedLinear
|
||||
@@ -106,7 +106,7 @@ class VisualTokenizer(torch.nn.Module):
|
||||
quant_config: QuantizationConfig | None = None,
|
||||
prefix: str = "",
|
||||
use_data_parallel: bool = False,
|
||||
attn_backend_override: _Backend | None = None,
|
||||
attn_backend_override: AttentionBackendEnum | None = None,
|
||||
):
|
||||
super().__init__()
|
||||
self.config = config
|
||||
@@ -135,7 +135,7 @@ class VisualTokenizer(torch.nn.Module):
|
||||
quant_config: QuantizationConfig | None = None,
|
||||
prefix: str = "",
|
||||
use_data_parallel: bool = False,
|
||||
attn_backend_override: _Backend | None = None,
|
||||
attn_backend_override: AttentionBackendEnum | None = None,
|
||||
):
|
||||
model_type = config.model_type
|
||||
if model_type == "siglip2_navit":
|
||||
|
||||
Reference in New Issue
Block a user