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

[bugfix] fix type[AttentionBackend] bug in kv_connector_base_v1 (#30051)

Browse Source 
Signed-off-by: 01267596 <xiongkai123@cmbchina.com>
Co-authored-by: 01267596 <xiongkai123@cmbchina.com>
This commit is contained in:
kx
2025-12-06 15:11:31 +08:00
committed by GitHub
parent a238cbd89d
commit d6aeaddf4a
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/distributed/kv_transfer/kv_connector/v1/base.py
View File

@@ -239,7 +239,7 @@ class KVConnectorBase_V1(ABC):
return return
def register_cross_layers_kv_cache( def register_cross_layers_kv_cache(
self, kv_cache: torch.Tensor, attn_backend: type[AttentionBackend] self, kv_cache: torch.Tensor, attn_backend: type["AttentionBackend"]
): ):
""" """
Initialize with a single KV cache tensor used by all layers. Initialize with a single KV cache tensor used by all layers.