Module rusty_machine::learning::optim::grad_desc

Gradient Descent

Implementation of gradient descent algorithm. Module contains the struct GradientDesc which is instantiated within models implementing the Optimizable trait.

Currently standard batch gradient descent is the only implemented optimization algorithm but there is flexibility to introduce new algorithms and git them into the same scheme easily.

Structs

AdaGrad

Adaptive Gradient Descent

GradientDesc

Batch Gradient Descent algorithm

RMSProp

RMSProp

StochasticGD

Stochastic Gradient Descent algorithm.