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 11, 07:30 UTC
chores: adjust the attn register param order (#30688)
Signed-off-by: Hank <hcc.mayday@gmail.com>
Hank_ committed
Dec 17, 2025 at 19:58 UTC
6482e3895baa483fb30227648aa4721f09699cba
1 file changed
+1
-1
vllm/attention/backends/registry.py
+1
-1
@@ -201,8 +201,8 @@ _MAMBA_ATTN_OVERRIDES: dict[MambaAttentionBackendEnum, str] = {}
201
202
def register_backend(
203
backend: AttentionBackendEnum | MambaAttentionBackendEnum,
204
- is_mamba: bool = False,
204
class_path: str | None = None,
205
+ is_mamba: bool = False,
206
) -> Callable[[type], type]:
207
"""Register or override a backend implementation.
208