Neural Network Project 🧠

🧠 Dynamic Multi-Label Neural Network built from scratch using a Perceptron-based architecture. Includes a flexible NeuralNetwork class with backpropagation and a CrossValidator for K-Fold cross-validation. Designed for training models with multiple binary outputs (e.g. predicting 3 possible diseases from patient features).

A small, minimal neural network project. This README explains the repository layout, how to run the project, and quick notes for development — all in English with emojis 🎉.

Overview 🚀

This repository contains a simple Python entry script and a dependency manifest. It's intended as a starting point for experiments, training, or inference workflows.

Project Structure 📁

• index.py — Main entry point. Run this script to start training, evaluation, or inference. It typically contains dataset loading and the model loop.
• requirements.txt — Lists Python packages required to run the project. Install with pip install -r requirements.txt.
• README.md — This file (explains the structure and how to run).

If you add more modules, consider placing them in a src/ or models/ folder and updating this section accordingly.

Prerequisites ✅

• Python 3.8+ recommended
• pip available in your PATH

Installation & Run 🛠️

Follow these steps to create an environment, install dependencies, and run the project:

# Create virtual environment (cross-platform)
python -m venv env

# Windows (PowerShell)
env\Scripts\Activate.ps1

# macOS / Linux
source env/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the main script
python index.py

If requirements.txt is empty or missing packages, open index.py to see which libraries are imported and add them to requirements.txt.

Tips & Notes 💡

• If you need GPU support (for example with PyTorch), install the appropriate package variant for your CUDA version.
• Keep hyperparameters and dataset paths configurable inside index.py for easier experimentation.