ashwaZ/gradient-descent-visualizer

Gradient Descent Visualizer

This Streamlit project explains optimization by showing gradient descent move across a 3D loss surface and a 2D contour map.

Concept Explanation

Gradient descent is one of the core optimization algorithms in machine learning. At every step, it moves parameters in the direction that most quickly reduces loss, which is the negative gradient.

This app makes that idea visual:

• —The 3D surface shows the full objective function.
• —The contour plot shows equal-loss regions from above.
• —The animated path shows how iterative updates move toward lower loss.

What You Can Explore

• —Learning rate
• —Starting position
• —Number of optimization steps