3 projects
calibrated-bo
Calibrated Bayesian optimization on top of bayesian-gp-cvloss: composable conformal prediction (Weighted + Localized + Adaptive), single- and multi-objective acquisition (cUCB / cEI / cEHVI), q-batch greedy with fantasies, and a one-call BO loop with persistent ACI state.
bayesian-gp-cvloss
A Python package for Gaussian Process Regression with hyperparameter optimization using Hyperopt and cross-validation, focusing on optimizing cross-validated loss.
bit-collision-free-MF
Generate collision-free Morgan fingerprints with hashed or unfolded mode.
