Skip to main content

> ML_ARCHITECTURE // STATE-SPACE-MODEL-MAMBA_v1.0

Selective State-Space Model (Mamba / Mamba-2)

Sub-quadratic sequence architecture featuring time-varying selection parameters and hardware-aware scan algorithms, delivering constant-memory inference and linear scaling with sequence length.

State Space Modelstextaudiotime-series
Back to All Architectures

Architecture Overview

Sub-quadratic sequence architecture featuring time-varying selection parameters and hardware-aware scan algorithms, delivering constant-memory inference and linear scaling with sequence length.

Implementing Libraries

PyTorchLinux Foundation / PyTorch Foundation · v2.4.1
View Spec
Mamba (mamba-ssm)Albert Gu & Tri Dao / State Spaces Team · v2.2.4
View Spec

Seminal Papers

Mamba: Linear-Time Sequence Modeling with Selective State SpacesAlbert Gu, Tri Dao (2023) · arXiv preprint
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space DualityTri Dao, Albert Gu (2024) · International Conference on Machine Learning (ICML)
Architectural Limitations & Constraints
  • Requires compatible deep learning framework and hardware acceleration for efficient execution.