Customisable Feed Forward Neural Network:

---

Description:

Python implementation of a Dense, Feed Forward Neural Network

• Command Line Interface to build the network

• Customise number of hidden layers in the network, and the activation function used per layer

• Plots the loss curve after n iterations (currently hardcoded to 5000)

---

Resources

• __main__.py - code which collates user input determining network structure

• training.py - code which runs feed forward and backprop for X no. of epochs

• ann_feed_forward.py - code which runs the feed forward loop through the network

• ann_backprop.py - code which updates the weights matrices

---

How to Use:

• python __main__.py
• Input the number of hidden layers in the network
• Input the activation function to be used
• Receive a printout of the loss curve for training and validation sets