AI / ML VisualizerRuns in your browser · No API keyOpen Model
← All topics

Training & Optimization

Weight Initialization

Compare zero, small, large, Xavier and He initialization.

Gradient magnitude through layers (log₁₀)Gradient magnitude through layers (log₁₀)0-161.25-10.52.5-53.750.556log₁₀ |∂output/∂input|
Activations · 1 × 5
Chain rule: ∏ wₗ f′(zₗ)Input gradient = 3.818e-5

Continue exploring

Gradient Descent →Optimizer Comparison →Vanishing & Exploding Gradients →

No account required. Local images and model files stay in your browser. Educational models demonstrate mechanisms; they are not production predictors.