Skip to content

ju-baer/Synaptix

Repository files navigation

🧠 Synaptix

Visualize Neural Networks Like Never Before
Real-time, interactive exploration of how artificial neurons connect, learn, and evolve.


🖼️ Preview

Synaptix Demo

This is how Synaptix looks in action.

🚀 Overview

Synaptix is a dynamic web-based tool that allows you to visualize, interact with, and understand the training and behavior of neural networks in real-time. Whether you're a curious student, a CS educator, or a seasoned AI practitioner, Synaptix opens the black box of deep learning through vivid, step-by-step animations and intuitive explanations.

“What a neuron sees, Synaptix shows.”


🎯 Key Features

  • 🔄 Live Training Visualization: Watch forward pass, backpropagation, and weight updates as they happen.
  • 🧩 Modular Network Builder: Drag-and-drop layers, set activations, adjust sizes.
  • 📊 Activation & Gradient Maps: See how each layer transforms data.
  • ⏯️ Step-by-Step Debugger: Pause, rewind, and inspect each computation.
  • 🌈 Custom Data & Problems: Train on your own 2D datasets or choose built-in challenges (e.g., XOR, spiral).
  • 🧠 Explainer Mode: Toggle human-readable insights on loss functions, overfitting, or optimization.

🧰 Tech Stack

Frontend Backend (optional) Other Integrations
React.js Node.js / Flask (optional) TensorFlow.js
D3.js / Canvas Express / REST API Anime.js, Chart.js
TailwindCSS / CSS3 Firebase (for session save) MathJax (for math rendering)

🚀 Getting Started

🌐 Live Demo

Check this project out at https://synaptix-xi.vercel.app/

🧪Run locally:

git clone https://github.com/ju-baer/Synaptix
cd Synaptix
npm install
npm run dev

About

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published