stan_epochs()#

liesel.goose.warmup.stan_epochs(warmup_duration=1000, posterior_duration=1000, init_duration=75, term_duration=50, base_duration=25, thinning_posterior=1, thinning_warmup=1)[source]#

Sets up a list of EpochConfig’s.

This follows the Stan Development Team (Stan Reference Manual 2021, Chapter 15.2 [1]).

Parameters:

  • warmup_duration (int) – The number of warmup samples. (default: 1000)

  • posterior_duration (int) – The number of posterior samples. (default: 1000)

  • init_duration (int) – The number of samples in the initial fast adaptation epoch. (default: 75)

  • term_duration (int) – The number of samples in the final fast adaptation epoch. (default: 50)

  • base_duration (int) – The number of samples in the first slow adaptation epoch. (default: 25)

  • thinning_posterior (int) – Thinning applied in the posterior epoch. (default: 1)

  • thinning_warmup (int) – Thinning applied in each warmup. (default: 1)

Return type:

list[EpochConfig]

Warning

Kernels which rely on history tuning might not be able to deal with thinning during warmup.

References