🧠 Neural Network with Backpropagation: Decision Boundary Lab (2D)
2D companion to the 3D neural network: a hand-written multilayer perceptron trained with backpropagation, SGD/momentum/Adam, L2 decay, five data sets, live decision boundary, loss curves and four in-page numerical checks including a finite-difference gradient check.
⚙ Under the hood
2D companion to the 3D neural network: a hand-written multilayer perceptron trained with backpropagation, SGD/momentum/Adam, L2 decay, five data sets, live decision boundary, loss curves and a finite-difference gradient check.
neural networkbackpropagationmultilayer perceptrongradient checkdecision boundarymachine learningCanvas 2D
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install