Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

EinSort: Sorting is All We Need for Tensorizing LLM

About

Tensor networks provide efficient representations for compressing large neural networks. By carefully designing shapes and topologies, they can significantly reduce memory and computational costs. However, identifying implicit low-rank structures in large foundation models remains challenging due to their enormous scale and un-structured weight distributions. We propose an adaptive tensorization method that discovers inherent low-rank structure in a target tensor by index ordering. Experiments on weight and KV-cache compression demonstrate improved reconstruction quality compared to baselines.

Toshiaki Koike-Akino, Jing Liu, Ye Wang• 2026

Related benchmarks

TaskDatasetResultRank
Robot ManipulationLIBERO
Spatial Success93.5
90
Robot ManipulationLIBERO
Spatial Success Rate98
32
Robot ManipulationLIBERO Spatial Object Goal Long
Spatial Success Score96
26
Showing 3 of 3 rows

Other info

Follow for update