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

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
This commit is contained in:
Matthew Bonanni
2026-01-27 10:02:51 -05:00
committed by GitHub
parent 1f3a2c2944
commit a608b4c6c2
151 changed files with 585 additions and 527 deletions
+1 -1
View File
@@ -21,7 +21,6 @@ from tests.compile.fusion_test_utils import (
from tests.utils import flat_product
from tests.v1.attention.utils import BatchSpec, create_common_attn_metadata
from vllm._custom_ops import cutlass_scaled_fp4_mm, scaled_fp4_quant
from vllm.attention.layer import Attention
from vllm.compilation.fusion_attn import ATTN_OP, AttnFusionPass
from vllm.compilation.fx_utils import find_op_nodes
from vllm.compilation.matcher_utils import QUANT_OPS
@@ -40,6 +39,7 @@ from vllm.config import (
set_current_vllm_config,
)
from vllm.forward_context import get_forward_context, set_forward_context
from vllm.model_executor.layers.attention import Attention
from vllm.model_executor.layers.quantization.utils.quant_utils import (
QuantKey,
kFp8StaticTensorSym,
+1 -1
View File
@@ -5,7 +5,6 @@ import pytest
import torch
from tests.compile.backend import TestBackend
from vllm.attention.layer import Attention
from vllm.compilation.matcher_utils import FLASHINFER_ROTARY_OP, RMS_OP, ROTARY_OP
from vllm.compilation.noop_elimination import NoOpEliminationPass
from vllm.compilation.post_cleanup import PostCleanupPass
@@ -21,6 +20,7 @@ from vllm.config import (
VllmConfig,
set_current_vllm_config,
)
from vllm.model_executor.layers.attention import Attention
from vllm.model_executor.layers.layernorm import RMSNorm
from vllm.model_executor.layers.rotary_embedding import RotaryEmbedding
from vllm.platforms import current_platform
+1 -2
View File
@@ -9,8 +9,7 @@ import torch
from tests.kernels.allclose_default import get_default_atol, get_default_rtol
from tests.kernels.utils import opcheck
from vllm import _custom_ops as ops
from vllm.attention.layer import Attention
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
from vllm.model_executor.layers.attention import Attention, MMEncoderAttention
from vllm.platforms import current_platform
from vllm.utils.mem_utils import get_max_shared_memory_bytes
from vllm.utils.torch_utils import set_random_seed
+1 -1
View File
@@ -12,7 +12,7 @@ from unittest.mock import patch
import pytest
import torch
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
from vllm.model_executor.layers.attention import MMEncoderAttention
from vllm.platforms import current_platform
from vllm.platforms.cpu import CpuPlatform
from vllm.platforms.cuda import CudaPlatform
+1 -1
View File
@@ -5,7 +5,6 @@ import numpy as np
import pytest
import torch
from vllm.attention.layer import Attention
from vllm.config import (
AttentionConfig,
CacheConfig,
@@ -19,6 +18,7 @@ from vllm.distributed.parallel_state import (
init_distributed_environment,
initialize_model_parallel,
)
from vllm.model_executor.layers.attention import Attention
from vllm.model_executor.layers.mamba.mamba_mixer2 import MambaMixer2
from vllm.platforms import current_platform
from vllm.sampling_params import SamplingParams
+3 -3
View File
@@ -7,7 +7,7 @@ from vllm.v1.worker.utils import bind_kv_cache
def test_bind_kv_cache(default_vllm_config):
from vllm.attention.layer import Attention
from vllm.model_executor.layers.attention import Attention
ctx = {
"layers.0.self_attn": Attention(32, 128, 0.1, prefix="layers.0.self_attn"),
@@ -35,7 +35,7 @@ def test_bind_kv_cache(default_vllm_config):
def test_bind_kv_cache_non_attention(default_vllm_config):
from vllm.attention.layer import Attention
from vllm.model_executor.layers.attention import Attention
# example from Jamba PP=2
ctx = {
@@ -58,7 +58,7 @@ def test_bind_kv_cache_non_attention(default_vllm_config):
def test_bind_kv_cache_draft_model(default_vllm_config):
from vllm.attention.layer import Attention
from vllm.model_executor.layers.attention import Attention
layer_names = [
"model.layers.0.attn",