AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model

Shengkun Tang, Yaqing Wang, Zhenglun Kong, Tianchi Zhang, Yao Li, Caiwen Ding, Yanzhi Wang, Yi Liang, Dongkuan Xu

CVPR 2023 · Architectures and efficiency › RNNs and normalization · approach: Neural scaling laws · 34 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 modality decomposition of the encoder (separate image/text early exits) (ablation)MuE · SNLI-VE testaccuracy: 88.5 vs 75.0 without decompositionImproved
add modality decomposition of the encoder (ablation)MuE · MS COCO Karpathy testCIDEr: 137.0 vs 117.5 without decompositionImproved
add layer-wise task loss on every decoder layer (ablation)MuE · MS COCO Karpathy testCIDEr: 137.0 vs 117.9 without task lossImproved
add layer-wise task loss on every decoder layer (ablation)MuE · SNLI-VE testaccuracy: 87.0 vs 88.5 without task lossNo clear effect
multiple early exiting via layer-wise similarity across encoder and decoderOFA Base · SNLI-VE testexpected time reduction: -50% time at 88.5 accuracy vs OFABase 89.2Improved
multiple early exiting via layer-wise similarity across encoder and decoderOFA Base · MS COCO Karpathy testCIDEr: 137 at -40.2% time vs DeeCap 129 at -38%Improved
DeeBERT-style decoder-only early exitingOFA Base · SNLI-VE testaccuracy: 78.8 vs 89.2 full modelHurt
training with smaller static model OFATiny instead of early exitingOFA · SNLI-VE testaccuracy: 85.2 vs 89.2 full modelHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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