AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Scaling & Shifting Your Features: A New Baseline for Efficient Model Tuning

Dongze Lian, Zhou Daquan, Jiashi Feng, Xinchao Wang

NeurIPS 2022 · Architectures and efficiency › Pruning and efficient tuning · approach: Adapter-based parameter-efficient tuning

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
Scale and shift features of a frozen pretrained modelViT-B/16 · FGVC (mean)Top-1 accuracy: 88.54 -> 90.72Improved
Scale and shift features of a frozen pretrained modelViT-B/16 · VTAB-1k (mean)Top-1 accuracy: 65.57 -> 73.10Improved
Use feature scaling and shifting instead of full fine-tuningViT-B/16 · ImageNet-1KTop-1 accuracy: 83.58 -> 83.10Hurt
Scale and shift features of a frozen pretrained modelViT-B/16 fine-tuned on ImageNet-1K · ImageNet-ATop-1 accuracy: 34.49 -> 45.88Improved
Scale and shift features of a frozen pretrained modelViT-B/16 fine-tuned on ImageNet-1K · ImageNet-RTop-1 accuracy: 51.29 -> 56.77Improved
Scale and shift features of a frozen pretrained modelViT-B/16 fine-tuned on ImageNet-1K · ImageNet-CmCE: 46.47 -> 41.47Improved
Use SSF with Swin-T instead of full fine-tuningSwin-T · COCO val2017 (Mask R-CNN)AP^b: 43.7 -> 34.9Hurt
Use SSF with Swin-T instead of full fine-tuningSwin-T · ADE20K (UPerNet)mIoU: 44.5 -> 38.9Hurt
Add SSF-ADA in more transformer layers (ablation)SSF on ViT-B/16 · CIFAR-100Top-1 accuracy: 88.70 -> 92.69 -> 93.30 -> 93.60 -> 93.99Improved
Add SSF-ADA after MLP operations (ablation)SSF on ViT-B/16 · CIFAR-100Top-1 accuracy: 93.46 -> 93.99Improved
Add SSF-ADA after attention operations (ablation)SSF on ViT-B/16 · CIFAR-100Top-1 accuracy: 93.69 -> 93.99Improved
Initialize scale factors around one and shift factors around zero rather than randomly around zero (ablation)SSF on ViT-B/16 · CIFAR-100Top-1 accuracy: 90.11 -> 93.99Improved

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