Basiс things to deploy the model on windows C++ in Visual Studio

Hi, I have a model (autoencoder) it works well in python. Now I need to embed it into the existing c++ project on Windows,
After reading the site, I haven’t yet the basic plan of how to do It. Do I have to Install TF2 with basel and use TF2 C++API ? I can’t find anything similar to .predict in python. Or do I have to use c_api? Can somebody explain to me the common practice?
Do I have to Use c_api, or C++ api? Or I have to use TFLite for inference?

See also: