AI Paper Atlas

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

MVP: Multimodality-Guided Visual Pre-training

Longhui Wei, Lingxi Xie, Wengang Zhou, Houqiang Li, Qi Tian

ECCV 2022 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 90 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
replace d-VAE tokenizer with CLIP vision branch as MIM target (MVP vs BEiT)BEiT ViT-B/16 · ImageNet-1K fine-tuningtop-1: 83.2 -> 84.4Improved
replace d-VAE tokenizer with CLIP vision branch (MVP vs BEiT)BEiT ViT-L/16 · ImageNet-1K fine-tuningtop-1: 85.2 -> 86.3Improved
replace d-VAE tokenizer with CLIP vision branch (MVP vs BEiT)BEiT ViT-B/16 · ADE20KmIoU: 45.6 -> 52.4Improved
replace d-VAE tokenizer with CLIP vision branch (MVP vs BEiT)BEiT ViT-B/16 · ImageNet-1K linear probingtop-1: 37.6 -> 75.4Improved
use CLIP features as target instead of HOG (MVP vs MaskFeat)ViT-B/16 · ImageNet-1K fine-tuningtop-1: 84.0 -> 84.4Improved
pre-train 300 epochs instead of 100 (ablation)MVP ViT-B/16 · ImageNet-1K fine-tuningtop-1: 83.9 -> 84.4Improved
pre-train 300 epochs instead of 100 (ablation)MVP ViT-B/16 · ADE20KmIoU: 52.0 -> 52.4Improved
use CLIP (multimodal) guidance instead of DINO features as target (ablation)MIM ViT-B/16 · ADE20KmIoU: 47.0 -> 52.4Improved
use CLIP (multimodal) guidance instead of DINO features as target (ablation)MIM ViT-B/16 · ImageNet-1K fine-tuningtop-1: 83.6 -> 84.4Improved
MVP on ImageNet-1K vs BEiT pre-trained on larger ImageNet-21KViT-B/16 · ADE20KmIoU: 46.3 -> 52.4Improved
MVP on ImageNet-1K vs BEiT pre-trained on ImageNet-21KViT-L/16 · ADE20KmIoU: 51.8 -> 54.3Improved
MVP linear probing vs contrastive DINOViT-B/16 · ImageNet-1K linear probingtop-1: 78.2 -> 75.4Hurt

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