← 返回论文库

Mamba: Linear-Time Sequence Modeling with Selective State Spaces

Gu, Dao (CMU + Princeton) · 2023
L4.1 · Foundation Model Tech StackarXiv:2312.00752#architecture#ssm
CORE IDEA
Selective SSM：sub-quadratic 序列建模，挑战 Transformer 在长序列任务上的统治。
L-ANCHOR · 为什么在这一层重要
post-Transformer 候选
arXiv:2312.00752 ↗源码 ↗

相关论文

Attention is All You Need
L4.12017
Scaling Laws for Neural Language Models
L4.12020
Training Compute-Optimal Large Language Models (Chinchilla)
L4.12022
DeepSeek-V3 Technical Report
L4.12024