Struct DropoutLayer 

pub struct DropoutLayer { /* private fields */ }

Dropout layer for regularization, backed by a seeded generator so training runs are reproducible.

Implementations

impl DropoutLayer

pub fn new(dropout_rate: f64) -> Self

Create a new dropout layer with a default seed. Starts in inference mode.

pub fn with_seed(dropout_rate: f64, seed: u64) -> Self

Create a new dropout layer with an explicit seed. Starts in inference mode.

pub fn forward<T: Float + Debug + ScalarOperand + Send + Sync + 'static>( &mut self, input: &Array2<T>, ) -> Array2<T>

Inverted dropout: zero entries with probability dropout_rate and scale the survivors by 1 / (1 - rate) so the expectation is preserved.

pub fn set_training(&mut self, training: bool)

Set training mode

pub fn is_training(&self) -> bool

Get training mode

pub fn dropout_rate(&self) -> f64

Dropout probability

Trait Implementations

impl Clone for DropoutLayer

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DropoutLayer

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

Formats the value using the given formatter.