Hands-On Deep Learning with Go

ML for Go

A library that has a more general collection of useful features is GoLearn. While DL-specific features are on its wish list, it has the necessary primitives to implement simple neural networks, random forests, clustering, and other ML approaches. It relies heavily on Gonum, a library that provides implementations of float64 and complex128 matrix structures and linear algebra operations on them.

Let's look at what this means from a code perspective, as shown here:

type Network struct {       origWeights *mat.Dense       weights *mat.Dense // n * n       biases []float64 // n for each neuron       funcs []NeuralFunction // for each neuron       size int       input int   }

Here, we have GoLearn's primary definition of what a neural network looks like. It contains definitions ...