The Definitive Guide toAI Data Centers
Ask the GuideAboutAccount
GuideGlossarySGD

SGD · stochastic gradient descent

The iterative optimization loop underlying model training; whether it synchronizes every step or only intermittently determines how far apart training sites can physically sit.

Current numbers

~500xless inter-site communication for DiLoCo (8 workers) vs fully-synchronous SGD, matching convergence on C4as of 2024 · register ↗

← All terms