Closed dflemin3 closed 4 years ago
This PR reverts some solver changes in the previous push to master (turns out the HODLR solver isn't the best for approxposterior) and updates both the documentation and GP hyperparameter optimization schemes.
This PR reverts some solver changes in the previous push to master (turns out the HODLR solver isn't the best for approxposterior) and updates both the documentation and GP hyperparameter optimization schemes.