Microsoft.ML.OnnxRuntime.Gpu 1.8.0: Failed to load library libonnxruntime_providers_shared.so

[asvyazin]

Hello, I'm using precompiled nuget package Microsoft.ML.OnnxRuntime.Gpu in .Net application on linux

Version 1.7.1 worked fine

Version 1.8.0 started throwing exception:

Microsoft.ML.OnnxRuntime.OnnxRuntimeException: [ErrorCode:Fail] OrtSessionOptionsAppendExecutionProvider_Cuda: Failed to load shared library
   at Microsoft.ML.OnnxRuntime.NativeApiStatus.VerifySuccess(IntPtr nativeStatus)
   at Microsoft.ML.OnnxRuntime.SessionOptions.AppendExecutionProvider_CUDA(Int32 deviceId)

It also logs this error before the exception:

2021-06-15 11:57:25.542258449 [E:onnxruntime:CSharpOnnxRuntime, provider_bridge_ort.cc:902 Ensure] Failed to load library libonnxruntime_providers_shared.so with error: libonnxruntime_providers_shared.so: cannot open shared object file: No such file or directory

I downloaded and unpacked nupkg file from nuget.org and it looks to me that libonnxruntime_providers_bridge.so is missing in linux-x64 folder:

Probably I'm missing something, any ideas?

System information

OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Docker Container nvidia/cuda:11.0.3-cudnn8-runtime-ubuntu20.04
ONNX Runtime installed from (source or binary): binary
ONNX Runtime version: 1.8.0
Python version: N/A
Visual Studio version (if applicable): N/A
GCC/Compiler version (if compiling from source): N/A
CUDA/cuDNN version: 11.0.3/8.x.x (not sure about exact cuDNN version)
GPU model and memory: nvidia A100 

[jywu-msft]

looks like it is missing the required .so files under linux-x64/native

[hendrik-skubch]

I run into the same issue when using the C++ API and call Ort::SessionOptions::AppendExecutionProvider_CUDA on linux.

[snnn]

Thank you. We are solving it.

[hariharans29]

Should be fixed by #8066