AI Paper Atlas

Research map / Architectures and efficiency / Quantization

REx: Data-Free Residual Quantization Error Expansion

Edouard YVINEC, Arnaud Dapogny, Matthieu Cord, Kevin Bailly

NeurIPS 2023 · Architectures and efficiency › Quantization · approach: Mixed-precision quantization · 3 citations (OpenAlex)

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 REx residual expansion (W4/A6 + 50% sparse residue) vs SPIQ at W6/A6MobileNet v2 · ImageNettop-1 accuracy: 63.24 -> 64.20Improved
add REx vs SPIQ at W6/A6EfficientNet B0 · ImageNettop-1 accuracy: 54.51 -> 57.63Improved
add REx vs SPIQ at W6/A6ResNet-50 · ImageNettop-1 accuracy: 75.98 -> 76.01Improved
add REx residue to DFQ quantizationOPT-13B · common sense reasoning averageaverage score: 51.47 -> 53.76Improved
add REx residue to SQuant quantizationOPT-13B · common sense reasoning averageaverage score: 51.33 -> 53.91Improved
add REx residue to PowerQuantOPT-13B · common sense reasoning averageaverage score: 54.38 -> 54.81Improved
add REx outlier residue to DFQOPT-13B · BoolQaccuracy: 54.71 -> 65.54Improved
add REx to PowerQuantOPT-13B · HellaSwagaccuracy: 51.29 -> 50.98Hurt
add REx to PowerQuantOPT-13B · BoolQaccuracy: 70.43 -> 69.45Hurt
add REx to DFQOPT-13B · OpenBookQAaccuracy: 25.80 -> 25.40Hurt
add 25% sparse residue to BrecQ in W4/A4 (ablation)MobileNet v2 · ImageNettop-1 accuracy: 66.57 -> 70.94Improved
increase sparse residue from 25% to 75% with SQuant (ablation)MobileNet v2 · ImageNet W4/A4top-1 accuracy: 58.64 -> 71.74Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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