AI Paper Atlas

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

TaiSu: A 166M Large-scale High-Quality Dataset for Chinese Vision-Language Pre-training

Yulong Liu, Guibo Zhu, Bin Zhu, Qi Song, Guojing Ge, Haoran Chen, GuanHui Qiao, Ru Peng, Lingxiang Wu, Jinqiao Wang

NeurIPS 2022 · Vision-language and multimodal › CLIP and zero-shot · approach: Self-distillation for vision-language pretraining · 1 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
filter web data by image-text retrieval (ablation)TaiSu ViT-B · Flickr8k retrievalMean Recall: 53.8 -> 58.2Improved
filter web data by image-text retrieval (ablation)TaiSu ViT-B · Flickr30k retrievalMean Recall: 58.7 -> 64.2Improved
filter web data by image-text retrieval (ablation)TaiSu ViT-B · COCO-CN retrievalMean Recall: 63.7 -> 66.4Improved
filter web data by image-text retrieval (ablation)TaiSu ViT-B · MUGE retrievalMean Recall: 48.8 -> 51.4Improved
filter web data by image-text retrieval (ablation)TaiSu ViT-B · zero-shot classification averageR@1: 48.13 -> 50.64Improved
filter web data by image-text retrieval (ablation)TaiSu ViT-B · EuroSAT classificationR@1: 36.0 -> 29.5Hurt
add generated captions to filtered web data (ablation)TaiSu ViT-B · Flickr8k retrievalMean Recall: 58.2 -> 72.6Improved
add generated captions to filtered web data (ablation)TaiSu ViT-B · Flickr30k retrievalMean Recall: 64.2 -> 77.7Improved
add generated captions to filtered web data (ablation)TaiSu ViT-B · EuroSAT classificationR@1: 29.5 -> 52.6Improved
add generated captions to filtered web data (ablation)TaiSu ViT-B · CIFAR10 classificationR@1: 89.2 -> 85.5Hurt
add generated captions to filtered web data (ablation)TaiSu ViT-B · zero-shot classification averageR@1: 50.64 -> 53.10Improved
add generated captions to filtered web data (ablation)TaiSu RN101 · zero-shot classification averageR@1: 46.63 -> 48.13Mixed

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