This is a simple implementation of deep Gaussian Processes (only two layers) using MCMC for sampling. The algorithm is based on A. Sauer et al. (Technometrics, 2023). The top-level script is demo.py (run python demo.py). A C++ implementation is also provided in dgp_cpp folder.
The deepgp_variational_inference_tutorial (which uses variational inference method) is also provided as a complementary for the MCMC method (run python deepgp_variational_inference_tutorial.py).