Stochastic gradient descent optimizer

Stochastic gradient descent optimizer

Usage

mlx_optimizer_sgd(params, lr = 0.01)

Arguments

params

List of parameters (from mlx_parameters()).

lr

Learning rate.

Value

An optimizer object with a step() method.

See also

mlx.optimizers.SGD

Examples

set.seed(1)
model <- mlx_linear(2, 1, bias = FALSE)
opt <- mlx_optimizer_sgd(mlx_parameters(model), lr = 0.1)