AI Paper Atlas

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

Bootstrapping Vision-Language Learning with Decoupled Language Pre-training

Yiren Jian, Chongyang Gao, Soroush Vosoughi

NeurIPS 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Test-time adaptation of VLMs · 5 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
add P-Former alignment loss to BLIP-2 (4M pairs)BLIP-2 OPT2.7B · Zero-shot VQAv2 valaccuracy: 46.8 -> 52.6Improved
add P-Former alignment loss to BLIP-2 (4M pairs)BLIP-2 OPT2.7B · Zero-shot OK-VQAaccuracy: 25.9 -> 30.0Improved
add P-Former alignment loss to BLIP-2 (4M pairs)BLIP-2 OPT2.7B · Zero-shot GQA test-devaccuracy: 30.5 -> 34.0Improved
add P-Former alignment loss to BLIP-2 for fine-tuned captioningBLIP-2 OPT2.7B · COCO Karpathy testCIDEr: 140.4 -> 141.8Improved
add P-Former alignment loss to BLIP-2 for NoCaps zero-shotBLIP-2 OPT2.7B · NoCaps overallCIDEr: 111.9 -> 115.1Improved
add alignment loss to stage-1 retrieval pre-trainingBLIP-2 stage 1 · Flickr30K image-to-textR@1: 94.3 -> 93.7No clear effect
add alignment loss to stage-1 retrieval pre-trainingBLIP-2 stage 1 · MSCOCO image-to-textR@1: 78.4 -> 78.7No clear effect
add alignment loss only in stage 1 (w1=10, w2=0) (ablation)BLIP-2 OPT2.7B · Zero-shot VQAv2 valaccuracy: 46.8 -> 51.4Improved
add alignment loss only in stage 2 (w1=0, w2=100) (ablation)BLIP-2 OPT2.7B · Zero-shot VQAv2 valaccuracy: 46.8 -> 50.4Improved
use alignment loss in both stages rather than stage 1 only (ablation)BLIP-2 OPT2.7B · Zero-shot VQAv2 valaccuracy: 51.4 -> 52.6Improved
add P-Former alignment loss with Flan-T5XL as LLM (ablation)BLIP-2 Flan-T5XL · Zero-shot VQAv2 valaccuracy: 48.3 -> 54.9Improved
train P-Former on 4M VL sentences instead of 12M LAION sentences (ablation)BLIP-2 + P-Former · Zero-shot VQAv2 valaccuracy: 52.6 -> 51.7Hurt

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