Why it matters
Seq length shapes training limits. Understanding shapes design.
Advertisement
The architecture
Attention: O(N^2) FLOPs + memory (without FA).
FFN: O(N * hidden).
KV cache scales with N.
Advertisement
How it works end to end
FlashAttention: no memory blow-up.
Ring attention: split across GPUs.
Compute: 6ND FLOPs for training.