Trait autograd::gradient_descent::Optimizer

pub trait Optimizer {
    fn update(&mut self, node: &Tensor, param: &mut NdArray, grad: NdArray);
}

Trait for any gradient descent optimizer

Required Methods

fn update(&mut self, node: &Tensor, param: &mut NdArray, grad: NdArray)

Updates the variable tensor

Updates param with grad. node is a symbolic representation of param

Implementors

• impl Optimizer for SGD
• impl Optimizer for Adam