AI Paper Atlas

Research map / Optimization / Optimizers

Win: Weight-Decay-Integrated Nesterov Acceleration for Faster Network Training

Pan Zhou, Xingyu Xie, Zhouchen Lin, Kim-Chuan Toh, Shuicheng Yan

JMLR 2024 · Optimization › Optimizers · approach: Adam convergence and adaptive learning rates

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 Win accelerationAdamW · ImageNet ResNet18top-1 accuracy (%): 67.9 -> 71.0Improved
add Win2 (three-step) accelerationLAMB · ImageNet ResNet18top-1 accuracy (%): 68.5 -> 71.3Improved
add Win accelerationAdam · ImageNet ResNet18top-1 accuracy (%): 66.5 -> 69.3Improved
add Win accelerationLAMB · ImageNet ResNet50 (avg epochs)top-1 accuracy (%): 78.7 -> 79.4Improved
add Win accelerationAdamW · ImageNet ViT-B (avg epochs)top-1 accuracy (%): 80.7 -> 81.7Improved
add Win accelerationAdamW · COCO Mask R-CNN Swinbox AP average: 51.5 -> 51.7Improved
add Win accelerationAdamW · Penn Treebank LSTMtest perplexity: 67.0 -> 66.5Improved
add Win2 accelerationAdam · WikiText-103 Transformer-XLtest perplexity: 26.7 -> 25.0Improved
replace Nesterov acceleration with Win (ablation)AdamW · ImageNet ResNet18top-1 accuracy (%): 69.3 -> 71.0Improved
replace Nesterov acceleration with Win (ablation)AdamW · ImageNet ResNet50 (100 epochs)top-1 accuracy (%): 78.4 -> 78.0Hurt
increase update steps from q=3 to q=6 (ablation)AdamW-Win2 · ImageNet ResNet18top-1 accuracy (%): 71.2 -> 71.0No clear effect
increase update steps from q=3 to q=6 (ablation)AdamW-Win2 · ResNet18 peak GPU memory (M)peak memory: 8094 -> 8588Hurt

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