liambeguin/neuralnetwork

Add momentum to SGD

Closed this issue · 2 comments

since the Hessian technique seems way too costly, lets add a momentum parameter to our SGD

can be done using a running average.

nop'