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, 22:30 UTC
[bugfix] fix type[AttentionBackend] bug in kv_connector_base_v1 (#30051)
Signed-off-by: 01267596 <xiongkai123@cmbchina.com> Co-authored-by: 01267596 <xiongkai123@cmbchina.com>
kx committed
Dec 6, 2025 at 15:11 UTC
d6aeaddf4a6201e35ec89bcd4b3719e4e7293f1f
1 file changed
+1
-1
vllm/distributed/kv_transfer/kv_connector/v1/base.py
+1
-1
@@ -239,7 +239,7 @@ class KVConnectorBase_V1(ABC):
239
return
240
241
def register_cross_layers_kv_cache(
242
- self, kv_cache: torch.Tensor, attn_backend: type[AttentionBackend]
242
+ self, kv_cache: torch.Tensor, attn_backend: type["AttentionBackend"]
243
):
244
"""
245
Initialize with a single KV cache tensor used by all layers.