Learning Rate Visualizer

Visualize how learning rate affects gradient descent convergence

Select Learning Rate

Too Small

0.001

Small

0.01

Optimal

0.1

Large

0.5

Too Large

1.5

Optimization Function

Loss Function:

Custom Learning Rate: 0.1

Controls

Animation Speed: Normal

Loss Landscape & Path

Loss vs Iterations

Optimization Statistics

Learning Rate

0.1

Iterations

Current Loss

Initial Loss

Gradient Norm

Status

Ready

About Learning Rates

The learning rate (α) controls how big a step gradient descent takes in each iteration. It's one of the most important hyperparameters in machine learning optimization.

Too Small: Convergence is very slow. Many iterations needed. Safe but inefficient.
Optimal: Fast convergence with stable path. Reaches minimum efficiently.
Too Large: Overshoots minimum, oscillates, or diverges completely.

Tips: Watch how the optimization path changes with different learning rates. Notice how small rates take tiny steps while large rates jump around or explode. The loss curve shows convergence speed and stability.