Struct rusty_machine::learning::optim::fmincg::ConjugateGD

pub struct ConjugateGD {
    pub rho: f64,
    pub sig: f64,
    pub int: f64,
    pub ext: f64,
    pub max: usize,
    pub ratio: f64,
    pub iters: usize,
}

Conjugate Gradient Descent algorithm

Fields

rho: f64

Constant in the Wolfe-Powell conditions.

sig: f64

Constant in the Wolfe-Powell conditions.

int: f64

Don’t reevaluate within int of the limit of the current bracket.

ext: f64

Extrapolate max of ext times the current bracket.

max: usize

Max of max function evaluations per line search

ratio: f64

The maximum allowed slope ratio

iters: usize

The default number of max iterations.

Trait Implementations

impl Clone for ConjugateGD

fn clone(&self) -> ConjugateGD

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for ConjugateGD

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

Formats the value using the given formatter.

impl Default for ConjugateGD

The default Conjugate GD algorithm.

The defaults are:

• rho = 0.01
• sig = 0.5
• int = 0.1
• ext = 3
• max = 20
• ration = 100
• iters = 100

fn default() -> ConjugateGD

Returns the “default value” for a type.

impl<M: Optimizable> OptimAlgorithm<M> for ConjugateGD

fn optimize(
    &self,
    model: &M,
    start: &[f64],
    inputs: &M::Inputs,
    targets: &M::Targets
) -> Vec<f64>

Return the optimized parameter using gradient optimization.

impl Copy for ConjugateGD