patel-zeel/gp_practical_tips

Gaussian Processes Practical Tips

These repo contains a collection of experiments on benchmark datasets to validate the practical tips to improve the performance of Gaussian processes.

Tips

• A practical guide to Gaussian processes

Datasets

• UCI datasets (available on request)

Collaboration guidelines

General guidelines

• You can take up a particular issue, discuss doubts in the issue itself and submit a PR to resolve the issue.

Technical details

• We are going to work with hydra for configuration management. For motivation behind using it, I highly recommend watching this video.
• If needed, we should be able to spin off multiple jobs (experiment configs) at once using hydra's cli interface.
• We should log statistics such as training and testing time, memory usage etc. using hydra's logging interface.