[5/N][Attention] Finish eliminating `vllm/attention` folder (#32064)

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>

vllm/model_executor/models/persimmon.py

@@ -30,11 +30,11 @@ import torch
 from torch import nn
 from transformers import PersimmonConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import get_act_fn
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,
     QKVParallelLinear,

vllm/model_executor/models/phi.py

@@ -45,11 +45,11 @@ import torch
 from torch import nn
 from transformers import PhiConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import get_act_fn
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,
     QKVParallelLinear,

vllm/model_executor/models/phimoe.py

@@ -31,10 +31,10 @@ import torch
 from torch import nn
 from transformers.configuration_utils import PretrainedConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.fused_moe import FusedMoE
 from vllm.model_executor.layers.linear import (
     QKVParallelLinear,

vllm/model_executor/models/plamo2.py

@@ -9,7 +9,6 @@ import torch
 from torch import nn
 from transformers import PretrainedConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import VllmConfig, get_current_vllm_config
 from vllm.distributed import divide, get_tensor_model_parallel_world_size
@@ -17,6 +16,7 @@ from vllm.distributed.parallel_state import get_pp_group
 from vllm.forward_context import ForwardContext, get_forward_context
 from vllm.model_executor.custom_op import CustomOp
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,

vllm/model_executor/models/plamo3.py

@@ -10,12 +10,12 @@ import torch
 from torch import nn
 from transformers import PretrainedConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import VllmConfig
 from vllm.distributed import get_tensor_model_parallel_world_size
 from vllm.distributed.parallel_state import get_pp_group
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
     MergedColumnParallelLinear,

vllm/model_executor/models/qwen.py

@@ -16,11 +16,11 @@ import torch
 from torch import nn
 from transformers import PretrainedConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
     MergedColumnParallelLinear,

vllm/model_executor/models/qwen2.py

@@ -33,12 +33,12 @@ import torch
 from torch import nn
 from transformers import Qwen2Config
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
-from vllm.model_executor.layers.attention.encoder_only_attention import (
+from vllm.model_executor.layers.attention import (
+    Attention,
     EncoderOnlyAttention,
 )
 from vllm.model_executor.layers.layernorm import RMSNorm

vllm/model_executor/models/qwen2_5_vl.py

@@ -49,7 +49,7 @@ from vllm.distributed import utils as dist_utils
 from vllm.forward_context import set_forward_context
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import get_act_and_mul_fn
-from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
+from vllm.model_executor.layers.attention import MMEncoderAttention
 from vllm.model_executor.layers.conv import Conv3dLayer
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (

vllm/model_executor/models/qwen2_moe.py

@@ -34,12 +34,12 @@ import torch.nn.functional as F
 from torch import nn
 from transformers import Qwen2MoeConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.fused_moe import SharedFusedMoE
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (

vllm/model_executor/models/qwen2_vl.py

@@ -49,7 +49,7 @@ from vllm.distributed import parallel_state, tensor_model_parallel_all_gather
 from vllm.distributed import utils as dist_utils
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import QuickGELU
-from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
+from vllm.model_executor.layers.attention import MMEncoderAttention
 from vllm.model_executor.layers.conv import Conv3dLayer
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,

vllm/model_executor/models/qwen3.py

@@ -30,11 +30,11 @@ import torch
 from torch import nn
 from transformers import Qwen3Config
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.logger import init_logger
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import QKVParallelLinear, RowParallelLinear
 from vllm.model_executor.layers.logits_processor import LogitsProcessor

vllm/model_executor/models/qwen3_moe.py

@@ -32,7 +32,6 @@ import torch
 import torch.nn.functional as F
 from torch import nn
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
 from vllm.distributed import (
@@ -43,6 +42,7 @@ from vllm.distributed import (
 )
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.fused_moe import SharedFusedMoE
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (

vllm/model_executor/models/qwen3_next.py

@@ -10,7 +10,6 @@ from einops import rearrange
 from torch import nn
 from transformers.activations import ACT2FN
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import (
     CacheConfig,
@@ -29,6 +28,7 @@ from vllm.distributed import (
 )
 from vllm.forward_context import ForwardContext, get_forward_context
 from vllm.logger import init_logger
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.fla.ops import (
     chunk_gated_delta_rule,
     fused_recurrent_gated_delta_rule,

vllm/model_executor/models/seed_oss.py

@@ -30,12 +30,12 @@ import torch
 from torch import nn
 from transformers import PretrainedConfig as SeedOssConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
     MergedColumnParallelLinear,

vllm/model_executor/models/siglip.py

@@ -19,10 +19,10 @@ from vllm.config import VllmConfig
 from vllm.config.multimodal import BaseDummyOptions
 from vllm.distributed import divide, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import get_act_fn
-from vllm.model_executor.layers.attention.encoder_only_attention import (
+from vllm.model_executor.layers.attention import (
     EncoderOnlyAttention,
+    MMEncoderAttention,
 )
-from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
 from vllm.model_executor.layers.conv import Conv2dLayer
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,

vllm/model_executor/models/siglip2navit.py

@@ -13,7 +13,7 @@ from transformers.configuration_utils import PretrainedConfig
 
 from vllm.distributed import divide, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import get_act_fn
-from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
+from vllm.model_executor.layers.attention import MMEncoderAttention
 from vllm.model_executor.layers.conv import Conv2dLayer
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,

vllm/model_executor/models/solar.py

@@ -30,11 +30,11 @@ import torch
 from torch import nn
 from transformers import PretrainedConfig
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
     MergedColumnParallelLinear,

vllm/model_executor/models/stablelm.py

@@ -29,10 +29,10 @@ import torch
 from torch import nn
 from transformers import StableLmConfig
 
-from vllm.attention.layer import Attention
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.linear import (
     MergedColumnParallelLinear,
     QKVParallelLinear,

vllm/model_executor/models/starcoder2.py

@@ -28,11 +28,11 @@ import torch
 from torch import nn
 from transformers import Starcoder2Config
 
-from vllm.attention.layer import Attention
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import get_act_fn
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.linear import (
     ColumnParallelLinear,
     QKVParallelLinear,

vllm/model_executor/models/step1.py

@@ -10,7 +10,6 @@ from collections.abc import Iterable
 import torch
 from torch import nn
 
-from vllm.attention.layer import Attention, AttentionType
 from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import (
     get_pp_group,
@@ -18,6 +17,7 @@ from vllm.distributed import (
     get_tensor_model_parallel_world_size,
 )
 from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.model_executor.layers.attention import Attention
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
     MergedColumnParallelLinear,
@@ -41,6 +41,7 @@ from vllm.model_executor.models.utils import (
     maybe_prefix,
 )
 from vllm.sequence import IntermediateTensors
+from vllm.v1.attention.backend import AttentionType
 
 STEP_PACKED_MODULES_MAPPING = {
     "qkv_proj": ["q_proj", "k_proj", "v_proj"],