AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Designing Energy-Efficient Convolutional Neural Networks Using Energy-Aware Pruning

Tien-Ju Yang, Yu-Hsin Chen, Vivienne Sze

CVPR 2017 · Architectures and efficiency › CNN design · approach: Compact CNN architectures · 732 citations (OpenAlex)

Check 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
Apply energy-aware pruningAlexNet · ImageNet ILSVRC 2014 validationNormalized energy per image: 3.97 -> 1.06Improved
Apply energy-aware pruning instead of magnitude-based pruningAlexNet · ImageNet ILSVRC 2014 validationNormalized energy per image: 1.85 -> 1.06Improved
Apply energy-aware pruningGoogLeNet · ImageNet ILSVRC 2014 validationNormalized energy per image: 7.63 -> 4.76Improved
Apply energy-aware pruningSqueezeNet · ImageNet ILSVRC 2014 validationNormalized energy per image: 5.28 -> 3.99Improved
Apply energy-aware pruning instead of magnitude-based pruningSqueezeNet · ImageNet ILSVRC 2014 validationNormalized energy per image: 4.61 -> 3.99Improved
Reduce the number of target classesAlexNet · ImageNet ILSVRC 2014 validation, reduced-class modelsNormalized energyMixed
Use model size as a proxy for energy consumptionCNN architecture comparison · ImageNet ILSVRC 2014Normalized energyHurt
Use number of MACs as a proxy for energy consumptionCNN architecture comparison · ImageNet ILSVRC 2014Normalized energyHurt

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