Convert PyTorch code to C++

I would like to convert some optimization (second-order non convex optimization) PyTorch code to C++. Ideally the code will also implement a supporting package which is imported and have CUDA support, but willing to fund work where the code uses an alternative optimizer like LBFGS and versions with no GPU parallelization support. Also interested in multiprecision support if possible.

Rate negotiable depending on ability to tend to details above.