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, 17:15 UTC
[Misc] Remove stale func in KVTransferConfig (#14746)
Signed-off-by: Shangming Cai <caishangming@linux.alibaba.com>
shangmingc committed
Mar 29, 2025 at 01:33 UTC
d03308be0c8d5c0e367bb67e8d6e158eb373f5e4
1 file changed
-6
vllm/config.py
-6
@@ -2986,12 +2986,6 @@ class KVTransferConfig(BaseModel):
2986
return self.kv_connector is not None and \
2987
self.kv_role in ["kv_producer", "kv_consumer", "kv_both"]
2988
2989
- @property
2990
- def need_kv_parallel_group(self) -> bool:
2991
- # for those database-based connector, vLLM does not need to create
2992
- # parallel group, and in that case the kv parallel size will be 1.
2993
- return self.kv_connector is not None and self.kv_parallel_size > 1
2994
-
2989
@property
2990
def is_kv_producer(self) -> bool:
2991
return self.kv_connector is not None and \