Why it matters
ZeRO enables large-model training. Understanding shapes memory optimization.
Advertisement
The architecture
Stage 1: shard optim states.
Stage 2: shard gradients too.
Stage 3: shard params too.
Advertisement
How it works end to end
Stage 1: 4x memory reduction.
Stage 2: 8x.
Stage 3: linear in N GPUs.
Comm cost rises with stage.