A Survey on Visual Mamba
State space models (SSM) with selection mechanisms and hardware-aware architectures, namely Mamba, have recently shown significant potential in long-sequence modeling. Since the complexity of transformers’ self-attention mechanism is quadratic with image size, as well as increasing computational dem...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
MDPI
2024
|