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 3, 14:04 UTC
[Model Runner V2] Fix kv_connector `pre_forward` order (#42676)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
May 15, 2026 at 11:13 UTC
af9616d845ed310f76687bc08ca928cdcb143232
1 file changed
+1
-1
vllm/v1/worker/gpu/kv_connector.py
+1
-1
@@ -65,8 +65,8 @@ class ActiveKVConnector(KVConnector):
65
66
kv_connector_metadata = scheduler_output.kv_connector_metadata
67
assert kv_connector_metadata is not None
68
- self.kv_connector.bind_connector_metadata(kv_connector_metadata)
68
self.kv_connector.handle_preemptions(kv_connector_metadata)
69
+ self.kv_connector.bind_connector_metadata(kv_connector_metadata)
70
71
# TODO: sort out KV Connectors' use of forward_context
72
if is_forward_context_available():