TuringLang / AdvancedHMC.jl

Robust, modular and efficient implementation of advanced Hamiltonian Monte Carlo algorithms
https://turinglang.org/AdvancedHMC.jl/
MIT License
228 stars 39 forks source link

kwarg nadapt #332

Closed JaimeRZP closed 1 year ago

JaimeRZP commented 1 year ago

The current constructors take n_adapts as well as other initialization options as a field to follow the Turing API as close as possible. However, ideally the intialization should be independent of the sampler. This PR is a small step towards that goal. In the future we would like to give other initialization options the same treatment.

yebai commented 1 year ago

Thanks, @JaimeRZP!