Python加载torchtrt_runtime.dll遇WinError1114问题求助

AI 小北

2026-8-20

问题：Python加载torchtrt_runtime.dll失败，无法加载torch_tensorRT模型

环境与现象

尝试在C++和Python中加载编译后的torch_tensorRT模块：

C++加载代码：

const torch::Device device = torch::Device(torch::kCUDA, 0);
model = torch::jit::load(model_path, device);

Python加载代码：

torch.jit.load(model_path)

C++ Debug模式运行正常，但Release模式和Python均报错：

RuntimeError: Unknown type name 'torch.torch.classes.tensorrt.Engine':
File "code/torch/movinets/models.py", line 4 parameters = [] buffers =
[]
_torch___movinets_models_MoViNet_trt_engine : torch.torch.classes.tensorrt.Engine < -- HERE def forward(self_1:
torch.movinets.models.MoViNet_trt, input_0: Tensor) -> Tensor:

已验证的有效操作与Python加载失败情况

在C++ Release模式中，通过预加载torchtrt_runtime.dll可正常加载模型，代码如下：

HMODULE hLib = LoadLibrary(TEXT("torchtrt_runtime"));
if (hLib == NULL) {
    std::cerr << "Library torchtrt_runtime.dll not found" << std::endl;
    exit(1);
}

但在Python中尝试以下两种方式加载torchtrt_runtime.dll均失败：

使用PyTorch内置方法：

torch.ops.load_library("torchtrt_runtime.dll")

使用ctypes：

import ctypes
hllDll = ctypes.WinDLL("torchtrt_runtime.dll") or ctypes.CDLL("torchtrt_runtime.dll")

报错详情

初始报错：

hllDll = ctypes.WinDLL ("torchtrt_runtime.dll") File
"C:\Users\NomanAnjum\AppData\Local\Programs\Python\Python310\lib\ctypes__init__.py",
line 374, in init self._handle = _dlopen(self._name, mode)
FileNotFoundError: Could not find module 'torchtrt_runtime.dll' (or
one of its dependencies). Try using the full path with constructor
syntax.

补全所有依赖DLL到同一目录后，报错变为：

File "video_play.py", line 198, in get_torch_tensorrt_converted_model
hllDll = ctypes.WinDLL(r"torchtrt_runtime.dll") File
"C:\Users\NomanAnjum\AppData\Local\Programs\Python\Python310\lib\ctypes__init__.py",
line 374, in init self._handle = _dlopen(self._name, mode)
OSError: [WinError 1114] A dynamic link library (DLL) initialization
routine failed