Metraipolis › archive › dc_2ae77e9db899

SSDi8: Accurate and Efficient 8-bit Quantization for State Space Duality — arXiv:2608.21952

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.21952
License (as archived): CC0-1.0
Archived: 2026-10-18T07:41:06.771Z
SHA-256 of the archived text: 161fd721008891e1c4c159eeea92e5324ed53ab9f55135a584d61a3fef268092
Stable URL of this page: https://metraipolis.com/documento/dc_2ae77e9db899

How to cite

SSDi8: Accurate and Efficient 8-bit Quantization for State Space Duality — arXiv:2608.21952 — arXiv (en), https://arxiv.org/abs/2608.21952 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2ae77e9db899 (Metraipolis archive dc_2ae77e9db899, sha256 161fd721008891e1c4c159eeea92e5324ed53ab9f55135a584d61a3fef268092)

BibTeX

@misc{dc2ae77e9db899,
  title = {SSDi8: Accurate and Efficient 8-bit Quantization for State Space Duality — arXiv:2608.21952},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.21952},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2ae77e9db899; sha256 161fd721008891e1c4c159eeea92e5324ed53ab9f55135a584d61a3fef268092},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.21952. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent advances in sequence modeling have highlighted Mamba as a state space architecture offering efficient long-range dependency modeling and providing a viable alternative to Transformers. Building upon this, Mamba-2 introduces the Structured State Space Duality (SSD), which integrates recurrent and attention modes to achieve efficiency and scalability. However, this architectural expansion substantially increases memory and latency overhead, underscoring the need for efficient compression strategies tailored to SSD. In this work, we present SSDi8, the first post-training quantization framework specifically designed for SSD to maintain a persistent INT8 path. SSDi8 introduces a reformulation that decouples element-wise multiplications from matrix multiplications, enabling reuse of quantized activations across modules. Moreover, SSDi8 adaptively quantizes channel-varying activations at cost-effective points, further reducing latency. On the accuracy side, SSDi8 explicitly leverages the intrinsic dimensional decomposition of SSD, exploiting distinct outlier distributions across axes, and incorporates an error correction term based on per-channel error statistics. Comprehensive experiments demonstrate that SSDi8 achieves accuracy comparable to FP16 while delivering up to 1.4x speedup in W4A8 and W8A8 settings. We further validate its robustness in resource-constrained environments by deploying it on the Orin NX device.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.21952 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T07:37:52.685Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).