AI Paper Atlas

Research map / Video / Video-language understanding

ST-Adapter: Parameter-Efficient Image-to-Video Transfer Learning

Junting Pan, Ziyi Lin, Xiatian Zhu, Jing Shao, Hongsheng Li

NeurIPS 2022 · Video › Video-language understanding · approach: Long video understanding · 76 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 3D depth-wise conv with temporal span (3x1x1) inside adapter vs 1x1x1 (ablation)ST-Adapter · SSv2 valtop-1 accuracy: 46.2 -> 66.3Improved
use spatial-only kernel (1x3x3) instead of temporal kernel (ablation)ST-Adapter · SSv2 valtop-1 accuracy: 46.2 -> 46.2No clear effect
kernel 3x3x3 vs 3x1x1 (ablation)ST-Adapter · K400 valtop-1 accuracy: 82.0 -> 82.0No clear effect
add adapters to deeper blocks (9-12) instead of shallow (1-4) (ablation)ST-Adapter · SSv2 valtop-1 accuracy: 45.9 -> 62.8Improved
add second ST-Adapter per block (before and after MHSA) (ablation)ST-Adapter · SSv2 valtop-1 accuracy: 65.6 -> 67.0Improved
add second ST-Adapter per block (before and after MHSA) (ablation)ST-Adapter · K400 valtop-1 accuracy: 82.0 -> 82.0No clear effect
increase bottleneck width from 64 to 384 (ablation)ST-Adapter · K400 valtop-1 accuracy: 81.4 -> 82.0Improved
add ST-Adapter to frozen CLIP ViT-B (vs full fine-tuning)CLIP ViT-B · SSv2 valtop-1 accuracy: 44.0 -> 66.3Improved
add ST-Adapter to frozen CLIP ViT-B/16CLIP ViT-B/16 · EK100 valverb top-1: 54.8 -> 67.6Improved
replace ST-Adapter with full fine-tuning on ImageNet-21K SSv2ViT-B IN21K · SSv2 valtop-1 accuracy: 64.4 -> 62.8Hurt
use ST-Adapter on Swin-B instead of full fine-tuningSwin-B · K400 valtop-1 accuracy: 81.5 -> 77.1Hurt
use ST-Adapter instead of standard Adapter (ablation)CLIP ViT-B/16 · SSv2 valtop-1 accuracy: 46.2 -> 66.3Improved

Later papers that compared with this method

6 later accepted papers report ST-Adapter as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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