Struct rusty_machine::learning::optim::grad_desc::RMSProp

pub struct RMSProp { /* fields omitted */ }

RMSProp

The RMSProp algorithm (Hinton et al. 2012).

Methods

impl RMSProp

fn new(learning_rate: f64,
       decay_rate: f64,
       epsilon: f64,
       iters: usize)
       -> RMSProp

Construct an RMSProp algorithm.

Requires learning rate, decay rate, epsilon, and iteration count.

Examples

use rusty_machine::learning::optim::grad_desc::RMSProp;

let rms = RMSProp::new(0.99, 0.01, 1e-5, 20);

Trait Implementations

impl Debug for RMSProp

fn fmt(&self, __arg_0: &mut Formatter) -> Result

Formats the value using the given formatter.

impl Clone for RMSProp

fn clone(&self) -> RMSProp

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Copy for RMSProp

impl Default for RMSProp

The default RMSProp configuration

The defaults are:

• learning_rate = 0.01
• decay_rate = 0.9
• epsilon = 1.0e-5
• iters = 50

fn default() -> RMSProp

Returns the "default value" for a type.

impl<M> OptimAlgorithm<M> for RMSProp where M: Optimizable<Inputs=Matrix<f64>, Targets=Matrix<f64>>

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

Return the optimized parameter using gradient optimization.