From 1c7bc1831808bf5e6d9b3283855d18951a1eb955 Mon Sep 17 00:00:00 2001 From: "Li, Jiang" Date: Mon, 22 Jun 2026 22:52:05 +0800 Subject: [PATCH] [Bugfix][CPU] Fix CPU model runner v2 (#46365) Signed-off-by: jiang1.li --- vllm/v1/worker/cpu/shm.py | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/vllm/v1/worker/cpu/shm.py b/vllm/v1/worker/cpu/shm.py index 9399b823ce..d691ada90b 100644 --- a/vllm/v1/worker/cpu/shm.py +++ b/vllm/v1/worker/cpu/shm.py @@ -17,6 +17,10 @@ def noop(*args: Any, **kwargs: Any) -> None: pass +def fake_pin_memory(self: torch.Tensor, *args: Any, **kwargs: Any) -> torch.Tensor: + return self + + class _EventPlaceholder: def __init__(self, *args, **kwargs) -> None: self.record = noop @@ -41,6 +45,7 @@ torch.cuda.set_stream = noop torch.cuda.current_stream = lambda *args, **kwargs: _StreamPlaceholder() torch.accelerator.synchronize = noop torch.accelerator.empty_cache = noop +torch.Tensor.pin_memory = fake_pin_memory # Patch vLLM torch utils import vllm.utils.torch_utils as torch_utils