Struct GradientDescent 

pub struct GradientDescent<Input, Output, M: OptimizableModel<Input, Output>> {
    pub cost_history: Vec<f64>,
    /* private fields */
}

A standard gradient descent optimizer.

Gradient descent is an optim algorithm that iteratively adjusts parameters to minimize a cost function by moving in the direction of the steepest decrease in the cost function.

Fields

• learning_rate - The step size for each iteration of gradient descent
• epochs - The number of complete passes through the training dataset
• cost_history - Records the cost value after each parameter update

Fields

cost_history: Vec<f64>

Implementations

impl<Input, Output, M: OptimizableModel<Input, Output>> GradientDescent<Input, Output, M>

pub fn new(learning_rate: f64, epochs: usize) -> Self

Creates a new GradientDescent optimizer.

Arguments

• learning_rate - The step size for each parameter update
• epochs - The number of complete passes through the training dataset

Returns

A new GradientDescent instance

Trait Implementations

impl<M: OptimizableModel<Matrix, Vector>> Optimizer<ArrayBase<OwnedRepr<f64>, Dim<[usize; 2]>>, ArrayBase<OwnedRepr<f64>, Dim<[usize; 1]>>, M> for GradientDescent<Matrix, Vector, M>

fn fit( &mut self, model: &mut M, x: &Matrix, y: &Vector, ) -> Result<(), ModelError>

Fits the model to the training data using gradient descent algorithm.

This method updates the model parameters by computing gradients and adjusting the parameters in the direction that minimizes the cost function.

Arguments

• model - The machine learning model to optimize
• x - The input training data
• y - The expected output values

Returns

• Ok(()) if optim completes successfully
• Err(ModelError) if an error occurs during optim