This repository contains a neural network implemented in numpy. It provides the following functionalities:
- feedforward neural network layer with bias and dropout
- ReLU or softmax for activation
- cross entropy or residual sum of squares for loss function
- layers can easily be stacked
- adjustible batch size
git clone https://github.com/davidtweedle/neuralnetwork.git
cd neuralnetwork
pip install .
To see a demonstration, you can check out the file example_mnist.ipynb