This is an interactive 3D neural network visualizer built with TensorFlow.js, Three.js, and React. Repo created by Zucker Timothy from Tanzania.
- 🔭 Visualizes neural networks in 3D with Three.js.
- 👨💻 Real time AI training.
- 🌍 Supports interactive learning for students and developers.
Frontend
Machine Learning
Tools
Neural Network Visualizer is a web app for exploring neural networks in 3D. It visualizes real time training and supports parameter tweaks for XOR, Linear Regression, and Circle Classification. This repo is a hobby of mine since i started to learn tensorflow.js
- 3D Visualization: View layers, weights, and activations in real time.
- Interactive Controls: Adjust layers, neurons, and learning rates.
- Problem Types: Train on XOR, Linear Regression, and Circle Classification.
- Metrics: Monitor real time performance and loss.
- Educational Focus: Enhances AI learning for Tanzanian and global audiences.
- Node.js 14+ and npm
- Git
-
Clone the repository:
git clone https://github.com/zuck30/neural-network-studio.git cd neural-network-studio -
Install dependencies:
npm install
-
Start the development server:
npm run dev
-
Access the app:
- Open http://localhost:5173 in your browser.
- Experiment with network parameters and visualize training.
- 3D Network Visualization: Animated weights and activations using Three.js.
- Real Time Training: Watch neural networks train in real-time.
- Adjustable Parameters: Customize layers, neurons, learning rate, and more.
- Problem Types: Supports XOR, Linear Regression, and Circle Classification.
- Training Stats: Displays metrics like loss and accuracy.
- Add more problems to visualize.
- Add more features.
- Consider desktop application.
We welcome contributions to enhance Neural Network Visualizer! See the CONTRIBUTING.md file for details.
This project is licensed under the MIT License. LICENSE.
For questions or issues, open a GitHub issue or contact mwalyangashadrack@gmail.com.