How to deploy tvm model for Cuda and OpenCL to create c++ API's

Are these questions the same as this one? Maybe someone could answer it in a centralized place?