Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality

Transformer是状态空间模型: 通过结构化状态空间对偶实现通用模型和高效算法。

Mamba-R: Vision Mamba ALSO Needs Registers

Mamba-R: 视觉Mamba也需要寄存器。

MambaOut: Do We Really Need Mamba for Vision?

MambaOut: 在视觉中我们真的需要Mamba吗？

Jamba: A Hybrid Transformer-Mamba Language Model

Jamba: 混合Transformer和Mamba的语言模型.

LoRA-GA: Low-Rank Adaptation with Gradient Approximation

LoRA-GA：梯度近似的低秩参数高效微调.

VMamba: Visual State Space Model

VMamba: 视觉状态空间模型.