Image Retrieval On Coco Cn
Metriken
R@1
R@10
R@5
Ergebnisse
Leistungsergebnisse verschiedener Modelle zu diesem Benchmark
| Paper Title | ||||
|---|---|---|---|---|
| CN-CLIP (ViT-H/14) | 81.5 | 99.1 | 96.9 | Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese |
| CN-CLIP (ViT-L/14@336px) | 80.1 | 99.2 | 96.7 | Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese |
| R2D2 (ViT-L/14) | 79.1 | 98.9 | 96.5 | CCMB: A Large-scale Chinese Cross-modal Benchmark |
| CN-CLIP (ViT-L/14) | 78.9 | 99.0 | 96.3 | Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese |
| CN-CLIP (ViT-B/16) | 77.0 | 99.0 | 97.1 | Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese |
| R2D2 (ViT-B) | 75.1 | 98.1 | 94.2 | CCMB: A Large-scale Chinese Cross-modal Benchmark |
| Wukong (ViT-L/14) | 74.0 | 98.1 | 94.4 | Wukong: A 100 Million Large-scale Chinese Cross-modal Pre-training Benchmark |
| Wukong (ViT-B/32) | 67.0 | 96.7 | 91.4 | Wukong: A 100 Million Large-scale Chinese Cross-modal Pre-training Benchmark |
| CN-CLIP (RN50) | 66.8 | 97.0 | 91.1 | Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese |
0 of 9 row(s) selected.