# Model-Building Basics

## Taking Gradients

Flux's core feature is taking gradients of Julia code. The `gradient`

function takes another Julia function `f`

and a set of arguments, and returns the gradient with respect to each argument. (It's a good idea to try pasting these examples in the Julia terminal.)

```
julia> using Flux.Tracker
julia> f(x) = 3x^2 + 2x + 1;
julia> df(x) = Tracker.gradient(f, x; nest = true)[1]; # df/dx = 6x + 2
julia> df(2)
14.0 (tracked)
julia> d2f(x) = Tracker.gradient(df, x; nest = true)[1]; # d²f/dx² = 6
julia> d2f(2)
6.0 (tracked)
```

(We'll learn more about why these numbers show up as `(tracked)`

below.)

When a function has many parameters, we can pass them all in explicitly:

```
julia> f(W, b, x) = W * x + b;
julia> Tracker.gradient(f, 2, 3, 4)
(4.0 (tracked), 1.0 (tracked), 2.0 (tracked))
```

But machine learning models can have *hundreds* of parameters! Flux offers a nice way to handle this. We can tell Flux to treat something as a parameter via `param`

. Then we can collect these together and tell `gradient`

to collect the gradients of all `params`

at once.

```
julia> using Flux
julia> W = param(2)
2.0 (tracked)
julia> b = param(3)
3.0 (tracked)
julia> f(x) = W * x + b;
julia> grads = Tracker.gradient(() -> f(4), params(W, b));
julia> grads[W]
4.0 (tracked)
julia> grads[b]
1.0 (tracked)
```

There are a few things to notice here. Firstly, `W`

and `b`

now show up as *tracked*. Tracked things behave like normal numbers or arrays, but keep records of everything you do with them, allowing Flux to calculate their gradients. `gradient`

takes a zero-argument function; no arguments are necessary because the `params`

tell it what to differentiate.

This will come in really handy when dealing with big, complicated models. For now, though, let's start with something simple.

## Simple Models

Consider a simple linear regression, which tries to predict an output array `y`

from an input `x`

.

```
W = rand(2, 5)
b = rand(2)
predict(x) = W*x .+ b
function loss(x, y)
ŷ = predict(x)
sum((y .- ŷ).^2)
end
x, y = rand(5), rand(2) # Dummy data
loss(x, y) # ~ 3
```

To improve the prediction we can take the gradients of `W`

and `b`

with respect to the loss and perform gradient descent. Let's tell Flux that `W`

and `b`

are parameters, just like we did above.

```
using Flux.Tracker
W = param(W)
b = param(b)
gs = Tracker.gradient(() -> loss(x, y), params(W, b))
```

Now that we have gradients, we can pull them out and update `W`

to train the model. The `update!(W, Δ)`

function applies `W = W + Δ`

, which we can use for gradient descent.

```
using Flux.Tracker: update!
Δ = gs[W]
# Update the parameter and reset the gradient
update!(W, -0.1Δ)
loss(x, y) # ~ 2.5
```

The loss has decreased a little, meaning that our prediction `x`

is closer to the target `y`

. If we have some data we can already try training the model.

All deep learning in Flux, however complex, is a simple generalisation of this example. Of course, models can *look* very different – they might have millions of parameters or complex control flow. Let's see how Flux handles more complex models.

## Building Layers

It's common to create more complex models than the linear regression above. For example, we might want to have two linear layers with a nonlinearity like sigmoid (`σ`

) in between them. In the above style we could write this as:

```
using Flux
W1 = param(rand(3, 5))
b1 = param(rand(3))
layer1(x) = W1 * x .+ b1
W2 = param(rand(2, 3))
b2 = param(rand(2))
layer2(x) = W2 * x .+ b2
model(x) = layer2(σ.(layer1(x)))
model(rand(5)) # => 2-element vector
```

This works but is fairly unwieldy, with a lot of repetition – especially as we add more layers. One way to factor this out is to create a function that returns linear layers.

```
function linear(in, out)
W = param(randn(out, in))
b = param(randn(out))
x -> W * x .+ b
end
linear1 = linear(5, 3) # we can access linear1.W etc
linear2 = linear(3, 2)
model(x) = linear2(σ.(linear1(x)))
model(rand(5)) # => 2-element vector
```

Another (equivalent) way is to create a struct that explicitly represents the affine layer.

```
struct Affine
W
b
end
Affine(in::Integer, out::Integer) =
Affine(param(randn(out, in)), param(randn(out)))
# Overload call, so the object can be used as a function
(m::Affine)(x) = m.W * x .+ m.b
a = Affine(10, 5)
a(rand(10)) # => 5-element vector
```

Congratulations! You just built the `Dense`

layer that comes with Flux. Flux has many interesting layers available, but they're all things you could have built yourself very easily.

(There is one small difference with `Dense`

– for convenience it also takes an activation function, like `Dense(10, 5, σ)`

.)

## Stacking It Up

It's pretty common to write models that look something like:

```
layer1 = Dense(10, 5, σ)
# ...
model(x) = layer3(layer2(layer1(x)))
```

For long chains, it might be a bit more intuitive to have a list of layers, like this:

```
using Flux
layers = [Dense(10, 5, σ), Dense(5, 2), softmax]
model(x) = foldl((x, m) -> m(x), layers, init = x)
model(rand(10)) # => 2-element vector
```

Handily, this is also provided for in Flux:

```
model2 = Chain(
Dense(10, 5, σ),
Dense(5, 2),
softmax)
model2(rand(10)) # => 2-element vector
```

This quickly starts to look like a high-level deep learning library; yet you can see how it falls out of simple abstractions, and we lose none of the power of Julia code.

A nice property of this approach is that because "models" are just functions (possibly with trainable parameters), you can also see this as simple function composition.

```
m = Dense(5, 2) ∘ Dense(10, 5, σ)
m(rand(10))
```

Likewise, `Chain`

will happily work with any Julia function.

```
m = Chain(x -> x^2, x -> x+1)
m(5) # => 26
```

## Layer helpers

Flux provides a set of helpers for custom layers, which you can enable by calling

`Flux.@treelike Affine`

This enables a useful extra set of functionality for our `Affine`

layer, such as collecting its parameters or moving it to the GPU.