For the installation on windows in C++ it seems we need to use a microsoft nuget package, available at https://www.nuget.org/packages/Microsoft.ML.OnnxRuntime.Gpu
Except that this package is basically empty: maybe it just provides the instructions to build from sources ?
Okay, so it seems the best path would be to build from the github repository, which is explained here: https://onnxruntime.ai/docs/build/training.html
(don't we really just have a package ready for download ?)
Ahh! Just found the list of assets on this release page: https://github.com/microsoft/onnxruntime/releases/tag/v1.17.1
Downloaded: