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 5, 14:45 UTC
[MRV2][KVConnector] Fix missing build_connector_worker_meta (#38698)
Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Yifan Qiao committed
Apr 2, 2026 at 22:42 UTC
311c9816470122fbe3b0a96820936f1a87955168
1 file changed
+4
vllm/v1/worker/gpu/kv_connector.py
+4
@@ -93,6 +93,10 @@ class ActiveKVConnector(KVConnector):
93
output.invalid_block_ids = self.kv_connector.get_block_ids_with_load_errors()
94
output.kv_connector_stats = self.kv_connector.get_kv_connector_stats()
95
output.kv_cache_events = self.kv_connector.get_kv_connector_kv_cache_events()
96
+ output.kv_connector_worker_meta = (
97
+ self.kv_connector.build_connector_worker_meta()
98
+ )
99
+
100
if clear_metadata:
101
self.kv_connector.clear_connector_metadata()
102
return output