This merge request creates a new mixin class which allows CUDA-specific models to be built more easily.