Mamba-2 introduced in: Dao & Gu 2024 — structured state space duality.
Object
Dao & Gu 2024 — structured state space duality
Primary source · preprint · 2024-05-31
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality — arXiv (Dao, Gu / Princeton + Carnegie Mellon)