Neural Network and activation function coded in Julia.

1/ Neural Network

This NN is a simple neural network coded in Julia. It contains a forward propagation function and a backpropagation function.

2/ Activation function

This project contains different activation functions, here is the list :

• Sigmoid
• TanH
• ReLU
• PReLU
• Softmax
• ELU
• Identity
• Binary

These functions are also derived, so these different functions are easily usable in an NN

3/ Usage

To use the NN, you need to download Julia.
So you will have to link your function to add features (an example function is given). Finally you can add a function to plot the data, like Plots or Gadfly.

In itself, this NN is not directly usable, it is more a kind of demonstration of the mathematical aspect of an NN.

Activation functions and derived functions can be easily added to a project.

4/ Links

Julia : https://julialang.org/
Blog : https://aidri.github.io/emping/