diff --git a/vllm/distributed/kv_transfer/infinite.py b/vllm/distributed/kv_transfer/infinite.py index 3055b45ffbaad..6a9c1f749b7e6 100644 --- a/vllm/distributed/kv_transfer/infinite.py +++ b/vllm/distributed/kv_transfer/infinite.py @@ -38,7 +38,7 @@ def __init__(self, model: str, tokens_per_page=16) -> None: logger.info("connecting to infinite store server: ", infinite_server) - self.conn.connect(infinite_server) + self.conn.connect() def _compute_kv_cache_block_offsets( self, input_ids: torch.Tensor, attn_metadata: AttentionMetadata,