diff --git a/vllm/lora/punica.py b/vllm/lora/punica.py index 6d5c834299961..d666fc293757b 100644 --- a/vllm/lora/punica.py +++ b/vllm/lora/punica.py @@ -10,8 +10,10 @@ import torch from vllm.triton_utils import HAS_TRITON +from vllm.utils import is_xpu -if HAS_TRITON: +# FIXME: xpu path doesn't support torch.library.custom_op +if HAS_TRITON and not is_xpu(): from vllm.lora.ops.bgmv_expand import bgmv_expand from vllm.lora.ops.bgmv_expand_slice import bgmv_expand_slice from vllm.lora.ops.bgmv_shrink import bgmv_shrink