Move visualization of GP output to a separate file

[acxz]

• Figure out how to load trained GP model checkpoint (need to make sure the original training data is saved with the rest of the model hyperparameters)

• This can also involve saving as an onnx file and loading it with tensorrt as in a production environment.

• I think to do this I need to first revamp how I am using hparams: See: https://pytorch-lightning.readthedocs.io/en/stable/hyperparameters.html?highlight=hyperparameter

[acxz]

Also adding a 2D visualization and 3D input-output case would be sooo cool. See tubes now thats cray cray

[acxz]

Something important to visualize is the overlayed Gaussians on the z axis at every point for the confidence region.