This is a Python implementation of a neural network built from scratch, using only NumPy for numerical computations. The network was also tested to classify images from the MNIST dataset, achieving an accuracy of over 96% on the validation set.
The purpose of this project is to provide a clear and concise implementation of the fundamental concepts behind neural networks, including feedforward and backpropagation algorithms, weight initialization, activation functions, and optimization techniques such as stochastic gradient descent.
This implementation is intended as a learning resource for anyone interested in understanding the inner workings of neural networks, and as a starting point for more complex implementations in real-world applications.
To get started with this project, simply clone this repository and run the whole jupyter NN_numpy.ipynb file. The script will download the MNIST dataset automatically and train the neural network for several epochs, printing out the training progress and validation accuracy.
This project requires the following Python packages:
- NumPy
- matplotlib (for visualization)
This project is licensed under the MIT License - see the LICENSE file for details.