Trait LossFunction 

pub trait LossFunction<T: FloatBounds + ScalarOperand> {
    // Required methods
    fn compute_loss(
        &self,
        predictions: &Array2<T>,
        targets: &Array2<T>,
    ) -> Result<T, SklearsError>;
    fn compute_gradient(
        &self,
        predictions: &Array2<T>,
        targets: &Array2<T>,
    ) -> Result<Array2<T>, SklearsError>;
}

Loss function trait for gradient checking

Required Methods

fn compute_loss( &self, predictions: &Array2<T>, targets: &Array2<T>, ) -> Result<T, SklearsError>

Compute loss given predictions and targets

fn compute_gradient( &self, predictions: &Array2<T>, targets: &Array2<T>, ) -> Result<Array2<T>, SklearsError>

Compute loss gradient with respect to predictions

Implementors

impl<T: FloatBounds + ScalarOperand> LossFunction<T> for CrossEntropyLoss<T>

impl<T: FloatBounds + ScalarOperand> LossFunction<T> for MeanSquaredError<T>