Home▸AI & Machine Learning▸Neural Network Architectures

🧠 Neural Network Architectures

A guided overview of neural network architectures, from feedforward networks to CNNs, RNNs and transformers.

AI & Machine Learning3DAdvanced60 FPS

A 3D layered network of neurons and connections that rebuilds its wiring live as you switch between feedforward, convolutional, recurrent and transformer architectures.

🔬 What It Demonstrates

Each architecture is defined by how its units are connected: dense layer-to-layer links for feedforward nets, local sliding windows for CNNs, chained hidden-state loops for RNNs, and all-to-all links for transformer attention.

🎮 How to Use

Pick an architecture from the dropdown, resize the network with the layer and neuron sliders, and watch glowing pulses travel along the active connections. Drag to rotate, scroll to zoom.

💡 Did You Know?

Weight sharing is what makes CNNs parameter-efficient — the same small filter slides across an entire image instead of learning a separate weight for every pixel pair.

⚙ Under the hood

A guided overview of neural network architectures, from feedforward networks to CNNs, RNNs and transformers.

artificial intelligenceneural networksdeep learningmachine learningalgorithmscomputer scienceThree.js

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install