Struct MultiStepLrScheduler 

pub struct MultiStepLrScheduler {
    pub initial_lr: f64,
    pub milestones: Vec<usize>,
    pub gamma: f64,
    /* private fields */
}

Multi-step learning rate scheduler.

Decays the learning rate by gamma at specified milestones (epochs). This is useful when you know specific points where you want to reduce LR.

Fields

initial_lr: f64

Initial learning rate.

milestones: Vec<usize>

Milestones (epochs) at which to decay LR.

gamma: f64

Multiplicative factor of learning rate decay.

Implementations

impl MultiStepLrScheduler

pub fn new(initial_lr: f64, milestones: Vec<usize>, gamma: f64) -> Self

Create a new multi-step LR scheduler.

Arguments

• initial_lr - Initial learning rate
• milestones - Epochs at which to decay (should be sorted)
• gamma - Multiplicative decay factor

Trait Implementations

impl Clone for MultiStepLrScheduler

fn clone(&self) -> MultiStepLrScheduler

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MultiStepLrScheduler

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

Formats the value using the given formatter.

impl LrScheduler for MultiStepLrScheduler

fn step(&mut self, optimizer: &mut dyn Optimizer)

Update learning rate based on current step/epoch.

fn get_lr(&self) -> f64

Get current learning rate.

fn state_dict(&self) -> HashMap<String, f64>

Get scheduler state as a dictionary.

fn load_state_dict(&mut self, state: &HashMap<String, f64>) -> TrainResult<()>

Load scheduler state from a dictionary.