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 15, 08:30 UTC
[MISC] typo fix and clean import (#18664)
Signed-off-by: Andy Xie <andy.xning@gmail.com>
Ning Xie committed
May 24, 2025 at 22:52 UTC
4ceafb6299c790ddc2be863a71227ae03da77c60
2 files changed
+2
-3
vllm/distributed/kv_transfer/__init__.py
+1
-2
@@ -1,8 +1,7 @@
1
# SPDX-License-Identifier: Apache-2.0
2
3
-from vllm.distributed.kv_transfer.kv_connector.base import KVConnectorBaseType
3
from vllm.distributed.kv_transfer.kv_transfer_state import (
5
- ensure_kv_transfer_initialized, get_kv_transfer_group,
4
+ KVConnectorBaseType, ensure_kv_transfer_initialized, get_kv_transfer_group,
5
has_kv_transfer_group, is_v1_kv_transfer_group)
6
7
__all__ = [
vllm/distributed/kv_transfer/kv_connector/simple_connector.py
+1
-1
@@ -106,7 +106,7 @@ class SimpleConnector(KVConnectorBase):
106
else:
107
108
# the current vLLM instance is KV consumer, so it needs to connect
109
- # its recv pipe to the send pipe of KV producder
109
+ # its recv pipe to the send pipe of KV producer
110
if self.config.kv_connector == "PyNcclConnector":
111
self.consumer_data_pipe = PyNcclPipe(
112
local_rank=local_rank,