AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Adaptive Nonlinear Compression for Large Foundation Models

Liang Xu, Shufan Shen, Qingming Huang, Yao Zhu, Xiangyang Ji, Shuhui Wang

ICLR 2026 · Architectures and efficiency › Quantization · approach: KV cache quantization

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add adaptive budget allocation (ablation)NLA on Swin-Tiny · ImageNet-1Ktop-1 accuracy (%): 77.40 -> 77.63Improved
add adaptive budget allocation (ablation)NLA on Swin-Base · ImageNet-1Ktop-1 accuracy (%): 82.55 -> 82.73Improved
add adaptive budget allocation (ablation)NLA on OPT-125M · WikiText-2perplexity: 31.52 -> 31.34Improved
add adaptive budget allocation (ablation)NLA on OPT-1.3B · WikiText-2perplexity: 20.13 -> 19.77Improved
use cubic instead of linear budget schedule (ablation)NLA on OPT-1.3B · WikiText-2perplexity: 43.24 -> 19.77Improved
use cubic instead of linear budget schedule (ablation)NLA on OPT-125M · WikiText-2perplexity: 63.04 -> 31.34Improved
use piecewise-linear kernel instead of Gaussian kernel (ablation)NLA on OPT-1.3B · inferencetime (s): 0.45 -> 0.25Improved
use piecewise-linear kernel instead of Gaussian kernel (ablation)NLA on LLaMA-7B · inferencetime (s): 1.45 -> 0.82Improved
use piecewise-linear kernel instead of Gaussian kernel (ablation)NLA on OPT-1.3B · peak GPU memorymemory (GB): 2.7 -> 2.8No clear effect
add memory-saving forward Algorithm 1 (ablation)NLA on OPT-1.3B · inference memorymemory (MB): 18053.07 -> 2204.22Improved
add memory-saving forward Algorithm 1 (ablation)NLA on OPT-1.3B · training memorymemory (MB): >81920 -> 12623.75Improved
use nonlinear NLA instead of linear SVD approximation (no training) (ablation)LLaMA-7B compression · WikiText-2perplexity: 32194 -> 16774Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy