Struct NutsSettings 

pub struct NutsSettings<A: Debug + Copy + Default + Serialize> {

    pub num_tune: u64,
    pub num_draws: u64,
    pub maxdepth: u64,
    pub mindepth: u64,
    pub store_gradient: bool,
    pub store_unconstrained: bool,
    pub store_transformed: bool,
    pub max_energy_error: f64,
    pub store_divergences: bool,
    pub adapt_options: A,
    pub check_turning: bool,
    pub target_integration_time: Option<f64>,
    pub trajectory_kind: KineticEnergyKind,
    pub num_chains: usize,
    pub seed: u64,
    pub extra_doublings: u64,
}

Settings for the NUTS sampler

Fields

num_tune: u64

The number of tuning steps, where we fit the step size and geometry.

num_draws: u64

The number of draws after tuning

maxdepth: u64

The maximum tree depth during sampling. The number of leapfrog steps is smaller than 2 ^ maxdepth.

mindepth: u64

The minimum tree depth during sampling. The number of leapfrog steps is larger than 2 ^ mindepth.

store_gradient: bool

Store the gradient in the SampleStats

store_unconstrained: bool

Store each unconstrained parameter vector in the sampler stats

store_transformed: bool

Store the transformed gradient and value in the sampler stats

max_energy_error: f64

If the energy error is larger than this threshold we treat the leapfrog step as a divergence.

store_divergences: bool

Store detailed information about each divergence in the sampler stats

adapt_options: A

Settings for geometry adaptation.

check_turning: bool

target_integration_time: Option<f64>

trajectory_kind: KineticEnergyKind

Selects the kinetic-energy form and the corresponding integrator.

• KineticEnergyKind::Euclidean: standard leapfrog (default for most settings).
• KineticEnergyKind::ExactNormal: geodesic leapfrog exact for a standard-normal potential.
• KineticEnergyKind::Microcanonical: isokinetic ESH-dynamics leapfrog (microcanonical HMC); momentum is constrained to the unit sphere.

num_chains: usize

seed: u64

extra_doublings: u64

Number of extra doublings to perform after reaching maxdepth. This can be used to increase the effective sample size at the cost of more expensive sampling.

Trait Implementations

impl<A: Clone + Debug + Copy + Default + Serialize> Clone for NutsSettings<A>

fn clone(&self) -> NutsSettings<A>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<A: Debug + Debug + Copy + Default + Serialize> Debug for NutsSettings<A>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de, A> Deserialize<'de> for NutsSettings<A>

where A: Deserialize<'de> + Debug + Copy + Default + Serialize,

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<A> Serialize for NutsSettings<A>

where A: Serialize + Debug + Copy + Default + Serialize,

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl<A: Copy + Debug + Copy + Default + Serialize> Copy for NutsSettings<A>