This is my implementation of a feedforward neural network with ReLU activation functions. I loosely followed the workflow outlined in the book https://nnfs.io.
The project has four main files:
NeuralNetwork.py
includes all the necessary classes for performing a forward and backward pass in the network.Optimizer.py
contains the stochastic gradient descent optimizer implementation and theTrainer
class responsible for training.Evaluation.py
includes the class that calculates the model'sAccuracy
.main.py
holds the code that trains and tests the neural network on a classification problem using spiral data.
Requirements are specified in requirements.txt
.