Home▸Physics & Mechanics▸Deep Neural Network Trainer: 2D Decision Map

🕸️ Deep Neural Network Trainer: 2D Decision Map

Train a small neural network live on 2D data and watch the decision map and weights change.

Physics & Mechanics2DModerate60 FPS📱 Mobile-adapted⇄ 2D version

Details

CategoryPhysics & Mechanics

RenderingCanvas 2D

DifficultyModerate

Target FPS60

ControlsSliders for hidden layers, neurons per layer, learning rate, L2 decay, batch size, epochs per frame, points and noise; choose dataset, input features, activation and optimizer; drag, add or double-click-delete data points, click neurons to switch them off, hover to probe activations; presets for circle, XOR, spiral, overfitting and regularisation.

deep learningneural networksbackpropagationdecision boundaryoverfittingAdam

Related simulations

SPH FluidFluid simulation via SPH: pressure, viscosity, gravity and wall bouncing. Splash impulse…

Spring-Mass Cloth SimulationExplore the physics of cloth movement with this simulation! Adjust the spring stiffness and…

Car PhysicsRealistic driving with suspension, drift, nitro and particle effects. Navigate rough terrain. ⊞ All 3207 in Physics & Mechanics →

⚙ Under the hood

Interactive 2D trainer for a small multilayer perceptron with live backpropagation: a decision-region map with draggable data points beside a network diagram whose edge colours show weight sign and size.

deep learningneural networksbackpropagationdecision boundaryoverfittingAdam

2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install