A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[CI/Build] Fix disabled v1 attention backend selection test (#25471)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Isotr0py committed
Sep 23, 2025 at 21:05 UTC
b6a136b58c10dde773045c471db762ed09fc7468
1 file changed
+1
-2
tests/kernels/attention/test_attention_selector.py
+1
-2
@@ -67,7 +67,6 @@ def generate_params():
67
return params
68
69
70
-@pytest.mark.skip(reason="Skipped for now. Should be revisited.")
70
@pytest.mark.parametrize("device, name, use_mla, block_size",
71
generate_params())
72
def test_env(
@@ -189,7 +188,7 @@ def test_env(
188
# FlashMLA only supports block_size == 64
189
pytest.skip("FlashMLA only supports block_size 64")
190
else:
192
- from vllm.attention.backends.flashmla import (
191
+ from vllm.v1.attention.backends.mla.flashmla import ( # noqa: E501
192
is_flashmla_supported)
193
is_supported, _ = is_flashmla_supported()
194
if not is_supported: