AI Paper Atlas

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

ELEVATER: A Benchmark and Toolkit for Evaluating Language-Augmented Visual Models

Chunyuan Li, Haotian Liu, Liunian Harold Li, Pengchuan Zhang, Jyoti Aneja, Jianwei Yang, Ping Jin, Houdong Hu, Zicheng Liu, Yong Jae Lee, Jianfeng Gao

NeurIPS 2022 · Vision-language and multimodal › CLIP and zero-shot · approach: Open-vocabulary alignment from VLMs · 63 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
initialize linear head from text embeddings (language-init) instead of random, linear probing (ablation)CLIP ViT-B32 · ELEVATER ICinW (20 datasets) 5-shotaccuracy: 58.09 -> 65.35Improved
language-init instead of random init, fine-tuning (ablation)CLIP ViT-B32 · ELEVATER ICinW (20 datasets) 5-shotaccuracy: 29.75 -> 63.29Improved
language-init, fine-tuning, full-shot (ablation)CLIP ViT-B32 · ELEVATER ICinW (20 datasets) full-shotaccuracy: 77.77 -> 80.35Improved
one-projection instead of two-projection language init (ablation)CLIP ViT-B32 · ELEVATER ICinW (20 datasets) 5-shotaccuracy: 65.35 -> 65.88Improved
CLIP language-augmented vs supervised ViT, 5-shot LPCLIP ViT-B32 · ELEVATER ICinW (20 datasets) 5-shotaccuracy: 56.00 -> 65.35Improved
fine-tuning vs linear probing, 5-shot, language-init (ablation)CLIP ViT-B32 · ELEVATER ICinW (20 datasets) 5-shotaccuracy: 65.35 -> 63.29Hurt
fine-tuning vs linear probing, full-shot (ablation)CLIP ViT-B32 · ELEVATER ICinW (20 datasets) full-shotaccuracy: 78.40 -> 80.35Improved
fine-tuning vs linear probing for detection (ablation)GLIP Swin-T · ELEVATER ODinW (35 datasets) 1-shotAP: 22.2 -> 32.2Improved
prompt tuning vs linear probing (ablation)GLIP Swin-T · ELEVATER ODinW (35 datasets) 1-shotAP: 22.2 -> 29.7Improved
add external knowledge (WordNet/Wiktionary/GPT3) to class text at zero-shot (ablation)UniCL/K-LITE · ELEVATER ICinW (20 datasets) zero-shotaccuracy: 27.15 -> 29.92~33.93Improved
add external knowledge for zero-shot detection (ablation)GLIP-A · ELEVATER ODinW (35 datasets) zero-shotAP: 11.53 -> 11.70~13.30Improved
more GPT3 knowledge items (ablation)K-LITE · IC vs OD zero-shotaccuracy / APMixed

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