Fleuret Explains Memory Trick That Cuts State Storage Complexity by a Dimension
Former Meta/Apple researcher François Fleuret shares a "wonderful trick" that reduces state storage complexity from bilinear O(timesteps × components) to O(timesteps), enabling both dimensions to scale, though with trade-offs.
2026-11-17 ~ 2026-11-17 · 2 related posts
- François Fleuret explains a memory trick: pay O(time steps) instead of O(steps × components) — francoisfleuret · 2026-11-17
- François Fleuret: clever O(T) memory trick, but it can't implement a real RNN — francoisfleuret · 2026-11-17