Struct RMSprop 

pub struct RMSprop { /* private fields */ }

RMSprop optimizer.

Maintains a moving average of squared gradients to normalize updates.

Update rule:

v_t = alpha * v_{t-1} + (1 - alpha) * grad^2
param = param - lr * grad / (sqrt(v_t) + eps)

With momentum:

v_t = alpha * v_{t-1} + (1 - alpha) * grad^2
buf_t = momentum * buf_{t-1} + grad / (sqrt(v_t) + eps)
param = param - lr * buf_t

Implementations

impl RMSprop

pub fn new(params: Vec<Parameter>, lr: f32) -> Self

Creates a new RMSprop optimizer with default settings.

pub fn with_alpha(params: Vec<Parameter>, lr: f32, alpha: f32) -> Self

Creates RMSprop with specified alpha (smoothing constant).

pub fn with_options( params: Vec<Parameter>, lr: f32, alpha: f32, eps: f32, weight_decay: f32, momentum: f32, centered: bool, ) -> Self

Creates RMSprop with all options.

pub fn alpha(self, alpha: f32) -> Self

Builder method to set alpha.

pub fn eps(self, eps: f32) -> Self

Builder method to set epsilon.

pub fn weight_decay(self, weight_decay: f32) -> Self

Builder method to set weight decay.

pub fn momentum(self, momentum: f32) -> Self

Builder method to set momentum.

pub fn centered(self, centered: bool) -> Self

Builder method to enable centered RMSprop.

Trait Implementations

impl Optimizer for RMSprop

fn step(&mut self)

Performs a single optimization step.

fn zero_grad(&mut self)

Zeros all parameter gradients.

fn get_lr(&self) -> f32

Returns the current learning rate.

fn set_lr(&mut self, lr: f32)

Sets the learning rate.

fn parameters(&self) -> &[Parameter]

Returns the parameters being optimized.

fn num_parameters(&self) -> usize

Returns the number of parameters.