This repository contains a neural network classification model implemented from scratch. The primary goal of this project is to provide a clear understanding of the operational logic behind neural networks.
- Implementation of a feedforward neural network architecture.
- Customizable number of layers and neurons per layer.
- Support for multiple activation functions (e.g., sigmoid, ReLU).
- Flexible training options, including customizable learning rate and batch size.
- Efficient backpropagation algorithm for updating network weights.
- Evaluation of model performance through accuracy metrics.
- Save and load trained models for future use.