Papers › ERNIE-ViL 2.0: Multi-view Contrastive Learning for Image-Text Pre-training

ERNIE-ViL 2.0: Multi-view Contrastive Learning for Image-Text Pre-training

30 Sep 2022arXiv:2209.15270archive 2025-07-28

Bin Shan, Weichong Yin, Yu Sun, Hao Tian, Hua Wu, Haifeng Wang

Recent Vision-Language Pre-trained (VLP) models based on dual encoder have attracted extensive attention from academia and industry due to their superior performance on various cross-modal tasks and high computational efficiency. They attempt to learn cross-modal representation using contrastive learning on image-text pairs, however, the built inter-modal correlations only rely on a single view for each modality. Actually, an image or a text contains various potential views, just as humans could capture a real-world scene via diverse descriptions or photos. In this paper, we propose ERNIE-ViL 2.0, a Multi-View Contrastive learning framework to build intra-modal and inter-modal correlations between diverse views simultaneously, aiming at learning a more robust cross-modal representation. Specifically, we construct multiple views within each modality to learn the intra-modal correlation for enhancing the single-modal representation. Besides the inherent visual/textual views, we construct sequences of object tags as a special textual view to narrow the cross-modal semantic gap on noisy image-text pairs. Pre-trained with 29M publicly available datasets, ERNIE-ViL 2.0 achieves competitive results on English cross-modal retrieval. Additionally, to generalize our method to Chinese cross-modal tasks, we train ERNIE-ViL 2.0 through scaling up the pre-training datasets to 1.5B Chinese image-text pairs, resulting in significant improvements compared to previous SOTA results on Chinese cross-modal retrieval. We release our pre-trained models in https://github.com/PaddlePaddle/ERNIE.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

PaddlePaddle/ERNIE officialmentioned in paperpaddle report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Computational EfficiencyContrastive LearningCross-Modal RetrievalImage RetrievalImage-to-Text RetrievalRetrievalZero-Shot Cross-Modal RetrievalZero-shot Image RetrievalZero-shot Text-to-Image Retrieval

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Image-to-text R@1 77.4 #17 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Image-to-text R@10 97.1 #17 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Image-to-text R@5 93.6 #17 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Text-to-image R@1 59.5 #17 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Text-to-image R@10 90.1 #17 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Text-to-image R@5 83.4 #17 of 36 Archive leaderboard report
Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Image-to-text R@1 97.2 #5 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Image-to-text R@10 100.0 #5 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Image-to-text R@5 100.0 #5 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Text-to-image R@1 93.3 #5 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Text-to-image R@10 99.8 #5 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Text-to-image R@5 99.4 #5 of 27 Archive leaderboard report
Image Retrieval AIC-ICC ERNIE-ViL2.0 Recall@1 19.0 #1 of 2 Archive leaderboard report
Image Retrieval AIC-ICC ERNIE-ViL2.0 Recall@10 43.5 #1 of 2 Archive leaderboard report
Image Retrieval AIC-ICC ERNIE-ViL2.0 Recall@5 35.3 #1 of 2 Archive leaderboard report
Image-to-Text Retrieval AIC-ICC ERNIE-ViL2.0 Recall@1 33.7 #1 of 2 Archive leaderboard report
Image-to-Text Retrieval AIC-ICC ERNIE-ViL2.0 Recall@10 60.0 #1 of 2 Archive leaderboard report
Image-to-Text Retrieval AIC-ICC ERNIE-ViL2.0 Recall@5 52.1 #1 of 2 Archive leaderboard report
Image-to-Text Retrieval Flickr30k ERNIE-ViL 2.0 Recall@1 96.1 #6 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k ERNIE-ViL 2.0 Recall@10 100.0 #6 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k ERNIE-ViL 2.0 Recall@5 99.9 #6 of 11 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Image-to-text R@1 63.1 #13 of 18 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Image-to-text R@10 91.4 #13 of 18 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Image-to-text R@5 85.7 #13 of 18 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Text-to-image R@1 46.0 #13 of 18 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Text-to-image R@10 80.4 #13 of 18 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval COCO 2014 ERNIE-ViL 2.0 Text-to-image R@5 71.4 #13 of 18 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Image-to-text R@1 91.2 #8 of 22 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Image-to-text R@10 99.8 #8 of 22 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Image-to-text R@5 99.1 #8 of 22 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Text-to-image R@1 77.4 #8 of 22 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Text-to-image R@10 96.4 #8 of 22 Archive leaderboard report
Zero-Shot Cross-Modal Retrieval Flickr30k ERNIE-ViL 2.0 Text-to-image R@5 93.8 #8 of 22 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Contrastive Learning

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections