Neural Network 3D Decision Boundary

NN

micrograd

Network Architecture

Input

Hidden

Hidden

Output

Architecture [3, 4, 4, 1]

Activation tanh

Parameters 37

Hyperparameters

Learning Rate: 0.01

Iterations: 200

Grid Resolution: 15

Render Every N Epochs: 1

Training Stats

0

Epoch

-

Loss

-

Accuracy

-

Time (ms)

Classification

Positive (y = 1)

Negative (y = -1)

Uncertain (conf < 0.25)

Changed (boundary shift)

Training Data (white ring)

Drag to rotate · Scroll to zoom · Shift+Drag to pan