inst/NEWS.md

In nimbleHMC: Hamiltonian Monte Carlo and Other Gradient-Based MCMC Sampling Algorithms for 'nimble'

CHANGES IN VERSION 0.2.3 (December 2024)

• Minor bug fix in the adapatation routines of the NUTS and NUTS_classic samplers, which could leave model in an inconsistent state.

CHANGES IN VERSION 0.2.1 (February 2024)

• Added checks within NUTS and NUTS_classic samplers to check validity of target nodes.

• Testing of correct results of HMC samplers on CAR model structures.

• NUTS and NUTS_classic samplers updated consistent with version 1.1.0 of nimble.

CHANGES IN VERSION 0.2.0 (September 2023)

• Renamed HMC sampler to NUTS_classic, and added NUTS HMC sampler which follows the implementation of version 2.32.2 of Stan.

• Replaced nwarmup control argument to HMC samplers with the new warmupMode control argument. This provides four distinct methods of choosing the number of HMC warmup iterations.

CHANGES IN VERSION 0.1.1 (July 2023)

• Fix bug in configureHMC for sampler assignment to posterior predictive nodes.

VERSION 0.1.0 (May 2023)

Initial release of nimbleHMC package. Initial version provides:

• Hamiltonian Monte Carlo (HMC) sampler for use with nimble package MCMC engine.

• Convenience functions addHMC, configureHMC, buildHMC and nimbleHMC, for managing MCMC configuration objects with HMC samplers.

nimbleHMC documentation built on April 4, 2025, 12:24 a.m.