AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Learning Rate Curriculum

Florinel-Alin Croitoru, Nicolae-Cătălin Ristea, Radu Tudor Ionescu, Nicu Sebe

IJCV 2024 · Architectures and efficiency › CNN design · approach: Compact CNN architectures · 39 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 LeRaC layer-wise learning rate curriculumResNet-18 · ImageNet (from scratch)accuracy: 68.44 -> 71.96Improved
add LeRaCResNet-18 · Food-101 (from scratch)accuracy: 68.31 -> 69.57Improved
add LeRaCCvT-13 (from scratch) · Food-101accuracy: 39.17 -> 41.42Improved
add LeRaCCvT-13 (pre-trained) · CIFAR-100accuracy: 77.80 -> 78.93Improved
add LeRaCBERT · BoolQaccuracy: 74.12 -> 75.55Improved
add LeRaCYOLOv5 · PASCAL VOC 2007+2012mAP: 0.832 -> 0.846Improved
add Curriculum by Smoothing (CBS)CvT-13 (pre-trained) · CIFAR-10accuracy: 93.56 -> 85.85Hurt
add CBSWide-ResNet-50 · Tiny ImageNet (from scratch)accuracy: 55.97 -> 48.30Hurt
add CBSResNet-18 · ImageNet (from scratch)accuracy: 68.44 -> 71.02Improved
use exponential instead of linear learning rate update in LeRaC (ablation)CvT-13 (from scratch) · CIFAR-10accuracy: 72.49 -> 72.90Improved
use exponential instead of logarithmic learning rate update in LeRaC (ablation)CvT-13 (from scratch) · CIFAR-10accuracy: 72.14 -> 72.90Improved
use curriculum (lower rates for deeper layers) instead of anti-LeRaC (ablation)ResNet-18 · CIFAR-100accuracy: 71.24 -> 72.72Improved

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