I am following this tutorial to implement a GP Regression using gPyTorch.
Based on my understanding of GP Regression, given the training data we can compute the posterior mean and covariance using the following formula,
But in the linked tutorial on gPyTorch documentation, there is a section dedicated to training the model across 50 iterations. From what I understand we only need to fit the data into the Gaussian process, and there is no need to iterate.
Can you help me understand?
My guess is that the tutorial is fitting different GP models with an RBF kernel of varying length scales. I could tell this only by looking at the print outputs of the training section.
Does that mean I can skip the training section if I was using a non-parameterized covariance kernel?
I tried looking for different gpytorch tutorials on the internet to see how people have implemented gpytorch with different kernels but could not find any. There were only a few re-hashes of the above tutorial.
