Glossary · Phase 4 · Deep learning
SGD
Gradient descent using a small random batch each step instead of all the data: noisier but much faster.
Try it now
A small batch gives many noisy steps per epoch. A large batch gives few smooth ones.
Learn it properly
Mini-batches and SGD · Train on small random batches instead of everything at once. About 30 minutes.
Never trained an AI?
Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account.