RandomFeatureGaussianProcess implementation for Deep Learning

Hi I need to implement this for school project:
[RandomFeatureGaussianProcess] (models/gaussian_process.py at master · tensorflow/models · GitHub)

It is based on using random fourier feature on gaussian process model that is end-to-end trainable with a deep neural network. (https://people.eecs.berkeley.edu/~brecht/papers/07.rah.rec.nips.pdf)

The github has some tensorflow codes but I hope to use pytorch. Can someone please show me some light on how I should be doing this?