Source: Dive into Deep Learning
In this example, we create a linear regression model that predicts housing data. It replicates the housing data example from the Knet.jl readme. Although we could have reused more of Flux (see the MNIST example), the library's abstractions are very lightweight and don't force you into any particular strategy.
A linear model can be created as a neural network with a single layer. The number of inputs is the same as the features that the data has. Each input is connected to a single output with no activation function. Then, the output of the model is a linear function that predicts unseen data.
To run this example:
cd other/housing/
julia --project housing.jl