AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

SDPT: Synchronous Dual Prompt Tuning for Visual-Language Pre-Trained Models

Yang Zhou, Yongjian Wu, Jiya Saiyin, Bingzheng Wei, Yan Xu

TPAMI 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Parameter-efficient visual tuning · 2 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
use shared unified prototype tokens instead of unshared tokens (ablation)SDPT · AquariummAP: 53.3 -> 57.7Improved
train both modalities synchronously instead of asynchronously (ablation)SDPT · AquariummAP: 54.8 -> 57.7Improved
use inverse linear projections instead of learnable projections (dual) (ablation)SDPT · AquariummAP: 54.3 -> 57.7Improved
apply dual-modal instead of image-only inverse projection (ablation)SDPT · AquariummAP: 53.1 -> 57.7Improved
apply dual-modal instead of text-only inverse projection (ablation)SDPT · AquariummAP: 55.4 -> 57.7Improved
vary token length k from 10 to 200 (ablation)SDPT · AquariummAP: 57.1 -> 57.7No clear effect
insert tokens in all 8 layers instead of layers 1-4 (ablation)SDPT · AquariummAP: 57.1 -> 57.7Improved
insert tokens in only layer 8 (ablation)SDPT · AquariummAP: 57.7 -> 57.5No clear effect
tune X-MHA with LoRA instead of encodersLoRA · COCOmAP: 56.5 -> 55.4Hurt
tune X-MHA with BitFit instead of encodersBitFit · COCOmAP: 54.5 -> 52.6Hurt
SDPT k=10 versus full fine-tuningGLIP-L · LVISmAP: 41.2 -> 41.2No clear effect
SDPT instead of new Adapter on frozen COCO adapterGLIP-L · COCO old-task mAPmAP: 35.4 -> 53.2Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

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