Thomas Christie

feat: Add some plotting functionality

1

# Feature Request Given how useful it is to visualise GPs for debugging, it would be nice to provide some basic plotting functionality out of the box (e.g. `plot_2d_gp`, `plot_1d_gp`)....

Currently when calling `predict` on a Gaussian process GPJax computes the entire covariance matrix. However, when calling with many test points this can be inefficient for downstream tasks which only...

[WIP] Add skeleton code for dealing with priors

1

Have added skeleton code for dealing with putting priors on kernel parameters. ## Type of changes - [ ] Bug fix - [X] New feature - [ ] Documentation /...

Problem training model with jax.jvp in forward pass

3

Hi - thanks for the library! I'm currently running into some issues when trying to train a "linearised" NN model. In my setup I'm taking a standard NN with weights...

Support for more low-rank curvature approximations

1

Hi - thanks for the library. I'm currently wanting to run some experiments using low-rank curvature approximations for regression. As far as I can tell, currently only the `AsdfghjklHessian` backend...