XAMBA makes Mamba and Mamba-2 state-space models run up to 4.8x faster on an Intel AI PC NPU by replacing CumSum and ReduceSum with matrix operations and activations with piecewise linear approximations.
HiPPO: Recurrent Memory with Optimal Polynomial Projections
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
XAMBA: Enabling Efficient State Space Models on Resource-Constrained Neural Processing Units
XAMBA makes Mamba and Mamba-2 state-space models run up to 4.8x faster on an Intel AI PC NPU by replacing CumSum and ReduceSum with matrix operations and activations with piecewise linear approximations.