Neural Networks

Program Descripiton:

This program has the following files: ml.cpp.
The program allows testing neural networks with n inputs and 1 input, with a hidden layer of variable size. The inputs can be given or generated by the program itself.
In addition to the program it is also included the code that generates inputs.

---

Requirements:

g++ compiler with c++ version 11.

---

How to compile and execute:

Compile: g++ -std=c++11 ml.cpp -o ml
Execute: ./ml

---

Execution:

At the beginning of the program the user will be asked for the number of inputs, number of neurons on the hidden layer, number of repetitions and learning rate. In each repetition the program gives an output: k, epochs (where k is the repetition number). If the model doesn't converge the number of epochs doesn't show up. At the end it is printed the average number of epochs, the number of repetitions that converged and the average time that it took to run 1000 epochs.

---

Autors:

Name	Numero UP
Diogo Pereira	201605323
Afonso Fernandes	201606852