AI Paper Atlas

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

Diverse Data Augmentation with Diffusions for Effective Test-time Prompt Tuning

Chun-Mei Feng, Kai Yu, Yong Liu, Salman Khan, Wangmeng Zuo

ICCV 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Parameter-efficient visual tuning · 118 citations (OpenAlex)

Read the paper

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
replace random-crop TPT augmentation with diffusion augmentation plus cosine filtration (DiffTPT)TPT on CLIP-RN50 · ImageNet-Atop-1 accuracy: 27.34 -> 31.06Improved
apply DiffTPT instead of TPTCLIP-RN50 OOD average · S1 OOD averagetop-1 accuracy: 43.46 -> 45.69Improved
apply DiffTPT on top of CoOp promptsCoOp (RN50) · S1 averagetop-1 accuracy: 47.12 -> 50.87Improved
apply TPT on top of CoCoOp with ViT-B/16CoCoOp (ViT-B/16) · S1 averagetop-1 accuracy: 61.01 -> 60.17Hurt
apply DiffTPT on top of CoCoOp with ViT-B/16CoCoOp (ViT-B/16) · S1 averagetop-1 accuracy: 61.01 -> 61.57Improved
use DiffTPT instead of TPT on cross-dataset generalizationTPT · S2 cross-dataset averageaverage top-1 accuracy: 55.12 -> 59.85 (RN50); 63.03 -> 65.47 (ViT-B/16)Improved
increase number of augmented views N up to 64 (ablation)DiffTPT · S1 and S2top-1 accuracyImproved
increase N to 128 (ablation)DiffTPT · S1 and S2top-1 accuracyNo clear effect
increase prompt update steps from 1 to 4 (ablation)DiffTPT · S2average top-1 accuracy: 56.6 -> 58.8Improved
increase prompt update steps beyond 5 (ablation)DiffTPT · S2average top-1 accuracyHurt
raise share of diffusion-based augmentation beyond ratio 0.4 (ablation)DiffTPT · S1top-1 accuracyHurt
TPT on EuroSAT versus zero-shot CLIPTPT · EuroSATtop-1 accuracy: reduces accuracy by 0.87 (RN50) and 3.75 (ViT-B/16)Hurt

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