MHKernel

class liesel.goose.MHKernel(position_keys, proposal_fn, initial_step_size=1.0, da_tune_step_size=False, da_target_accept=0.234, da_gamma=0.05, da_kappa=0.75, da_t0=10, identifier='')

Bases: ModelMixin, TransitionMixin[RWKernelState, DefaultTransitionInfo], ReprMixin

A Metropolis-Hastings kernel implementing the Kernel protocol.

The user needs to provide a proposal function that proposes a new state and the log_correction.

If da_tune_step_size is True the stepsize passed as an argument to the proposal function is tuned using the dual averging algorithm. Step size is tuned on the fly during all adaptive epochs.

Methods

end_epoch(prng_key, kernel_state, ...)	Sets the step size as found by the dual averaging algorithm.
end_warmup(prng_key, kernel_state, ...)	Currently does nothing.
init_state(prng_key, model_state)	Initializes the kernel state.
start_epoch(prng_key, kernel_state, ...)	Resets the state of the dual averaging algorithm.
tune(prng_key, kernel_state, model_state, epoch)	Currently does nothing.

Attributes

error_book	Dict of error codes and their meaning.
identifier	Kernel identifier, set by EngineBuilder
needs_history	Whether this kernel needs its history for tuning.
position_keys	Tuple of position keys handled by this kernel.