onnxruntime-gpu版本环境配置记录

•  onnx官方文档链接：Python | onnxruntime
• 环境配置

环境配置

显卡	RTX3060
显卡驱动	531.14
CUDA	NVCUDA64.DLL 为12.1.68	官方文件版本cuda_12.1.0_531.14_windows
cuDNN	8.9.7.29_cuda12
onnxruntime	1.18.0

•   安装onnxruntime：

  # onnxruntime推理环境，安装GPU，首先通过清华源安装，安装的同时会自动安装onnxruntimegpu的附属依赖，
  # pip install  onnxruntime-gpu==1.18.0 -i  https://pypi.tuna.tsinghua.edu.cn/simple
  # 然后卸载onnxruntimegpu，然后通过官网源安装一遍
  # pip install onnxruntime-gpu==1.18.0 -i https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

 pip install onnxruntime-gpu==1.18.0 -i https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

• 测试程序（路径更换成自己的模型路径）

  import onnxruntime
  
  print(onnxruntime.__version__)
  
  # print(onnxruntime.get_build_info())
  print(onnxruntime.get_device())
  print(onnxruntime.get_available_providers())
  
  
  modelpath = 'G:/YOLO-testPrj/onnxgpu_link/modelfile/yolo8n.onnx'
  MYprovider = ['CUDAExecutionProvider']
  session = onnxruntime.InferenceSession(path_or_bytes=modelpath, providers=MYprovider)
  print('use')
  print(session.get_providers())
  print(session.get_outputs())
  print(session.get_provider_options())