biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

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

Browse Source 
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
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/v1/worker/test_utils.py
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",