diff --git a/vllm/_xpu_ops.py b/vllm/_xpu_ops.py index 09f700d0de7..6a2e5e841ce 100644 --- a/vllm/_xpu_ops.py +++ b/vllm/_xpu_ops.py @@ -387,14 +387,9 @@ class xpu_ops: assert len(window_size) == 2 real_window_size = (window_size[0], window_size[1]) # noqa: F841 - # In encode attention, k and v maybe not contiguous and current - # kernel can't handle it - if block_table is None: - k = k.contiguous() - v = v.contiguous() return flash_attn_varlen_func( out=out, - q=q.contiguous(), + q=q, k=k, v=v, cu_seqlens_q=cu_seqlens_q,