onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be load

错误提示：

RuntimeError:D:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:891 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be loaded. Please install the correct version of CUDA andcuDNN as mentioned in the GPU requirements page  (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements),  make sure they're in the PATH, and that your GPU is supported.

1、根据这篇博客，确保cuda和cudnn的安装正确

win10下安装 Anaconda + Cuda + Cudnn + Pycharm + Pytorch_win10安装conda-CSDN博客

2、排除cuda是否正确安装问题，和确定cuda的版本

nvcc -V

​

3、版本cuda、cudnn和onnxruntime-gpu依赖关系,见官网Install ONNX Runtime | onnxruntime

cuda 版本为12.x时，请使用下面的方式进行重新安装

pip uninstall onnxruntime-gpu
pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

4、如果以上方法还不行，请执行安装命令

pip install onnxruntime-openvino

若出现拒绝访问Anaconda的环境时，根据提示可加参数 --user

pip install onnxruntime-openvino --user

5、最后个人开发环境在第4步，问题解决。

个人开发环境：cuda12.2 + cudnn8.9.2.26 + torch2.2.0