ML Optimizer Lab

Optimizers

Visualize how optimizers learn

Optimizers

Gradient Descent SGD Momentum Adam Muon

Loss Function

Hyperparameters

Learning Rate 0.0020

Speed 30

SGD Noise 0.15

Starting point
x = -1.50
y = 1.50

Click anywhere on the landscape to choose a new starting point.

Loss Landscape

x / y parameter space

Click to choose start

Gradient Descent

SGD

Momentum

Adam

Muon

Start

Minimum

Live Metrics

Step

0

Loss

-

X

-

Y

-

Optimizer

Gradient Descent

Select an optimizer and press Run.

Global minimum: x = 1, y = 1

Optimization Playground