Struct ExponentialStochasticDepth 

pub struct ExponentialStochasticDepth {
    pub num_layers: usize,
    pub drop_prob_min: f64,
    pub drop_prob_max: f64,
}

Exponential stochastic depth scheduler.

Exponentially increases drop probability across network depth. Provides more aggressive regularization in deeper layers.

Fields

num_layers: usize

Total number of layers/blocks.

drop_prob_min: f64

Drop probability for first layer.

drop_prob_max: f64

Drop probability for last layer.

Implementations

impl ExponentialStochasticDepth

pub fn new( num_layers: usize, drop_prob_min: f64, drop_prob_max: f64, ) -> TrainResult<Self>

Create a new exponential stochastic depth scheduler.

pub fn get_drop_prob(&self, layer_idx: usize) -> f64

Get drop probability for a specific layer using exponential interpolation.

pub fn create_drop_paths(&self) -> TrainResult<Vec<DropPath>>

Create DropPath instances for all layers.

Trait Implementations

impl Clone for ExponentialStochasticDepth

fn clone(&self) -> ExponentialStochasticDepth

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ExponentialStochasticDepth

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

Formats the value using the given formatter.