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 15, 08:30 UTC
[CI] fix kv_cache_type argument (#18594)
Signed-off-by: Andy Xie <andy.xning@gmail.com>
Ning Xie committed
May 23, 2025 at 19:49 UTC
cd821ea5d29f66ee0fd3a6c4580a089011e95343
1 file changed
+1
-1
tests/plugins_tests/test_platform_plugins.py
+1
-1
@@ -29,5 +29,5 @@ def test_oot_attention_backend(monkeypatch: pytest.MonkeyPatch):
29
# ignore the backend env variable if it is set
30
with monkeypatch.context() as m:
31
m.setenv(STR_BACKEND_ENV_VAR, STR_INVALID_VAL)
32
- backend = get_attn_backend(16, torch.float16, torch.float16, 16, False)
32
+ backend = get_attn_backend(16, torch.float16, "auto", 16, False)
33
assert backend.get_name() == "Dummy_Backend"