Search papers, labs, and topics across Lattice.
HAMSA, a novel vision state space model, eliminates the need for scanning by operating directly in the spectral domain via FFT-based convolutions. It introduces a simplified kernel parameterization, SpectralPulseNet for input-dependent frequency gating, and a Spectral Adaptive Gating Unit for stable training. HAMSA achieves state-of-the-art accuracy among SSMs on ImageNet-1K (85.7%) with significant improvements in inference speed, memory usage, and energy efficiency compared to transformers and scanning-based SSMs.
Vision models can ditch the scanning: HAMSA achieves state-of-the-art accuracy among SSMs while being 2.2x faster than transformers and using significantly less memory and energy.
Vision State Space Models (SSMs) like Vim, VMamba, and SiMBA rely on complex scanning strategies to adapt sequential SSMs to process 2D images, introducing computational overhead and architectural complexity. We propose HAMSA, a scanning-free SSM operating directly in the spectral domain. HAMSA introduces three key innovations: (1) simplified kernel parameterization-a single Gaussian-initialized complex kernel replacing traditional (A, B, C) matrices, eliminating discretization instabilities; (2) SpectralPulseNet (SPN)-an input-dependent frequency gating mechanism enabling adaptive spectral modulation; and (3) Spectral Adaptive Gating Unit (SAGU)-magnitude-based gating for stable gradient flow in the frequency domain. By leveraging FFT-based convolution, HAMSA eliminates sequential scanning while achieving O(L log L) complexity with superior simplicity and efficiency. On ImageNet-1K, HAMSA reaches 85.7% top-1 accuracy (state-of-the-art among SSMs), with 2.2 X faster inference than transformers (4.2ms vs 9.2ms for DeiT-S) and 1.4-1.9X speedup over scanning-based SSMs, while using less memory (2.1GB vs 3.2-4.5GB) and energy (12.5J vs 18-25J). HAMSA demonstrates strong generalization across transfer learning and dense prediction tasks.
