Skip to content
Hot eventLive

提出分布鲁棒量化DRQ改进低位宽LLM量化

1 reports1 sources3 hr ago updated

Get the story

AI overview

2026年10月9日,arXiv Statistics Machine Learning发布论文《低重构损失为何反而有害:面向低位 LLM 量化的分布鲁棒精炼》(一手报道)。论文提出Distributionally Robust Quantization(DRQ),是一种针对weight-only PTQ的后处理精炼框架,其核心思路是在受限输入激活分布集合上最小化最坏情况重构损失,以改进低位宽LLM量化效果。该报道同时提出“低重构损失为何反而有害”的问题视角,但目前仅见这一篇报道,尚无后续验证或第三方复现信息。

Generated from reports · updated 3 hr ago

Timeline

Follow the coverage from different angles.

Oct 9, 2026
  1. arXiv · Statistics Machine Learning
    低重构损失为何反而有害:面向低位 LLM 量化的分布鲁棒精炼

    论文提出 Distributionally Robust Quantization(DRQ),一种针对 weight-only PTQ 的后处理精炼框架,在受限输入激活分布集合上最小化最坏情况重构损失。

Heat trend

Current heat 9·Comparable peak 10(Oct 9)·Comparable change over 24 hours –

02.557.510Oct9Oct9Oct9Oct9

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.