Distributed Data Parallel and FSDP from Scratch
Multi-rank training is two collectives and one rule. Broadcast the parameters at startup, average the gradients after backward, never let the ranks disagree about what step they are on.
Visual edition planned
This lesson isn’t interactive yet.
It is part of the curriculum and will get the same treatment as Phase 1 — a visual cover, hands-on labs, derivations with numeric checks and a quiz. Until then, the original lesson is the best place to read it:
Part of Phase 19 — Capstone Projects. Use the previous / next cards below to keep browsing the phase.