AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Distillation-Based Training for Multi-Exit Architectures

Mary Phuong, Christoph H. Lampert

ICCV 2019 · Architectures and efficiency › RNNs and normalization · approach: Efficient network compression and training · 189 citations (OpenAlex)

Read the paper

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
train with distillation from the last exit instead of exit-wise loss onlyMSDNet · ImageNet(100) Exit 1top-5 accuracy (%): 64.4 -> 68.1Improved
train with distillation from the last exit instead of exit-wise loss onlyMSDNet · ImageNet subsetstop-5 accuracy (%): accuracy gains of up to 3.8%Improved
train with distillation from the last exit instead of exit-wise loss onlyMSDNet · CIFAR100 subsetstop-5 accuracy (%): up to 6.3% improvementImproved
train with distillation instead of exit-wise loss on full ImageNetMSDNet · ImageNet(full)top-5 accuracy (%): Exit 1: 87.8 -> 88.8; Exit 5: 92.3 -> 91.7Mixed
add unlabeled data to distillation (semi-supervised) (ablation)MSDNet · CIFAR100 Exit 1top-5 accuracy (%): 1.2%, 0.9%, and 0.4% increase for CIFAR(150), CIFAR(250), CIFAR(350)Improved
add unlabeled data to distillation (semi-supervised) (ablation)MSDNet · ImageNettop-5 accuracy (%): small additional improvement, e.g. ImageNet(100) Exit 5: 69.3 -> 69.7Improved
add unlabeled data to distillation (semi-supervised), anytime-mode (ablation)MSDNet · CIFAR(150) late exitstop-5 accuracy (%)Mixed
reduce amount of labeled training data (effect on distillation benefit)MSDNet · ImageNettop-5 accuracy (%): Exit 1: 64.4 -> 68.1 at ImageNet(100) vs 87.8 -> 88.8 at fullImproved
add adaptive temperature annealing instead of fixed temperature (ablation)distillation-trained MSDNet · CIFAR(250)top-5 accuracy (%)No clear effect
choose different teacher exits instead of the last exit (ablation)distillation-trained MSDNet · CIFAR100/ImageNettop-5 accuracy (%)No clear effect

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