HomeMachine Learning & Neural NetworksBatch Size Hyperparameters: Gradient Descent Simulator

📦 Batch Size Hyperparameters: Gradient Descent Simulator

Interactive 3D simulator of batch size hyperparameters: watch mini-batches of samples drive gradient descent across a loss landscape, and see how batch size trades off gradient noise, updates per epoch, and convergence stability.

Machine Learning & Neural Networks3DModerate60 FPS
batch-size-hyperparameters-complete-guide ↗ Open standalone
⚙ Under the hood

Watch a mini-batch of samples drive gradient descent across a bumpy loss landscape: shrink the batch size to see gradient noise and exploration rise, or grow it to see the path smooth out into fewer, more stable updates per epoch.

Three.jsmachine learningbatch sizegradient descenthyperparametersInstancedMesh

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install

What did you find?

Add reproduction steps (optional)