AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Image Captioners Are Scalable Vision Learners Too

Michael Tschannen, Manoj Kumar, Andreas Peter Steiner, Xiaohua Zhai, Neil Houlsby, Lucas Beyer

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Concept-guided image captioning · 10 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 contrastive pretraining (CLIP*) by CapPa captioning pretrainingViT-L/14 · COCO captioning with fresh decoderCIDEr: 123.2 -> 125.8Improved
replace CLIP* by CapPaViT-L/14 · VQAv2 with fresh decoderaccuracy: 68.5 -> 70.9Improved
replace CLIP* (16k) by CapPaViT-B/16 · ImageNet-1k 10-shot-style linear adaptation with decoderaccuracy: 81.4 -> 81.3No clear effect
replace CLIP* by CapPa for full ImageNet finetuningViT-L/14 · ImageNet-1k finetuningaccuracy: 88.1 -> 87.7Hurt
add parallel prediction to Cap (CapPa) (ablation)Cap · ImageNet 10-shot, 900M examplesaccuracy: 49.7 -> 50.4Improved
add parallel prediction to Cap (CapPa) (ablation)Cap · ImageNet-1k with decoderaccuracy: 80.2 -> 81.3Improved
randomly reverse captions in Cap (forward+backward) (ablation)Cap · ImageNet 10-shotaccuracy: 49.7 -> 49.2No clear effect
use two 3-layer decoders instead of one 6-layer decoder in CapPa (ablation)CapPa · ImageNet 10-shotaccuracy: 50.4 -> 49.5Hurt
use frozen T5 decoder instead of trained-from-scratch decoder in Cap (ablation)Cap · ImageNet 10-shotaccuracy: 48.7 -> 42.8Hurt
unfreeze T5 decoder (ablation)Cap with T5 decoder · ImageNet 10-shotaccuracy: 42.8 -> 48.6Improved
remove image input from Cap decoder (blind decoder) (ablation)Cap · ARO VG Attributionaccuracy: 88.9 -> 83.7Hurt
apply captioning to ViT-B/16 instead of ResNet50 encoder (ablation)Cap vs CLIP* · ImageNet zero-shotrelative accuracyImproved

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