https://github.com/iamfarrokhnejad/neural-network
A simple neural network without reliance on any NN library.
https://github.com/iamfarrokhnejad/neural-network
classification matplotlib mnist mnist-classification mnist-dataset mnist-handwriting-recognition neural-network neural-networks neuralnetwork neuralnetworks numpy pandas python
Last synced: 6 months ago
JSON representation
A simple neural network without reliance on any NN library.
- Host: GitHub
- URL: https://github.com/iamfarrokhnejad/neural-network
- Owner: IAmFarrokhnejad
- License: mit
- Created: 2025-01-02T00:06:01.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-01-19T21:21:13.000Z (9 months ago)
- Last Synced: 2025-02-03T03:34:46.341Z (8 months ago)
- Topics: classification, matplotlib, mnist, mnist-classification, mnist-dataset, mnist-handwriting-recognition, neural-network, neural-networks, neuralnetwork, neuralnetworks, numpy, pandas, python
- Language: Python
- Homepage:
- Size: 5.86 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Neural Network
## Overview
This repository contains the implementation of a neural network developed from scratch in Python without using any deep learning libraries (e.g., PyTorch, TensorFlow). The project showcases the process of building, training, and evaluating a custom neural network.
## Features
- Implementation of feedforward and backpropagation algorithms.
- Modular architecture with customizable activation functions, layers, and optimizers.
- Training accuracy of up to 92.63% over 2500 iterations.
## Prerequisites
Ensure you have the following installed on your system:
- Python 3.8+
- MNIST dataset: https://www.kaggle.com/datasets/oddrationale/mnist-in-csv
- Required libraries (install via `requirements.txt`):
```bash
pip install pandas
```
## Usage
### 1. Clone the Repository
```bash
git clone https://github.com/IAmFarrokhnejad/Neural-Network
cd your-repo-name
```
### 2. Prepare the Data
Place your dataset files in the `data/` directory. Ensure that the data is preprocessed as required by the training script.
### 3. Train the Neural Network
Run the training script to start the training process:
```bash
python mnist.py
```
Training logs, including iteration-wise accuracy, will be displayed in the console.
## Sample Output
Example of training accuracy progression:
```
Iteration 0, Accuracy: 0.1285
Iteration 1000, Accuracy: 0.9147
Iteration 2500, Accuracy: 0.9263
Iteration 7000, Accuracy: 0.9378
Iteration 9990, Accuracy: 0.9429
```
## Contributions
Contributions are welcome! Feel free to open an issue or submit a pull request.
## License
This project is licensed under the MIT License. See `LICENSE` for details.