🛰️ Daily AI Frontier
‹ back to 2026-09-12

Structured Transforms for Low-Overhead Quantization of Language Models

Research Efficiency & Systems

Ranking

Overall 78
Content 95
Popularity 37

Observed public metrics from 1 member.

Representative image for Structured Transforms for Low-Overhead Quantization of Language Models

Merged summary

TL;DR - Kashin-DCT is a language-model weight quantization method that replaces costly dense random transforms with sign-randomized DCTs, cutting each iteration from $\mathcal{O}(N^2)$ to $\mathcal{O}(N \log N)$. It enables efficient 4-bit-per-channel compression while improving numerical stability on configurations where competing methods fail.

  • A greedy alternating-update algorithm produces the four-peak distributions needed to cluster each factor into stable 2-bit codes.
  • Closed-form cluster-center initialization eliminates the multi-restart k-means bottleneck used by earlier approaches.
  • Combined with OPTQ-style error compensation and QuIP-style incoherence preprocessing, the JAX implementation is competitive on OPT, Llama-2, and Pythia models with favorable wall-clock scaling.
  • The bounded-$\ell_\infty$ factorization remains near FP16 perplexity in reported Pythia-6.9B and Mistral-7B stress tests where QuIP variants diverge or encounter NaNs.

Sources (1)

Structured Transforms for Low-Overhead Quantization of Language Models

arXiv cs.CL Daria Cherniuk, Alexander Rudikov, Boris Kashin, Ivan Oseledets 2026-09-10 arXiv:2609.11687
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-22 14:22:26.035364 UTC

TL;DR - Kashin-DCT is a language-model weight quantization method that replaces costly dense random transforms with sign-randomized DCTs, cutting each iteration from $\mathcal{O}(N^2)$ to $\mathcal{O}(N \log N)$. It enables efficient 4-bit-per-channel compression while improving numerical stability on configurations where competing methods fail.

  • A greedy alternating-update algorithm produces the four-peak distributions needed to cluster each factor into stable 2-bit codes.
  • Closed-form cluster-center initialization eliminates the multi-restart k-means bottleneck used by earlier approaches.
  • Combined with OPTQ-style error compensation and QuIP-style incoherence preprocessing, the JAX implementation is competitive on OPT, Llama-2, and Pythia models with favorable wall-clock scaling.
  • The bounded-$\ell_\infty$ factorization remains near FP16 perplexity in reported Pythia-6.9B and Mistral-7B stress tests where QuIP variants diverge or encounter NaNs.
item →