Cross-modal Retrieval on MS-COCO 5K
433.3rSumCRCL
Evaluation results
| Method | Links | |
|---|---|---|
| CRCLNoise Rate=20%, Evaluation Protocol=Fine-tune2023.10 | 433.3 | |
| CLIPNoise Rate=0%, Evaluation Protocol=Zero-Shot, Backbone=ViT-L/142023.10 | 400.4 | |
| CLIPNoise Rate=0%, Evaluation Protocol=Zero-Shot, Backbone=ViT-B/322023.10 | 361.6 | |
| CLIPNoise Rate=20%, Evaluation Protocol=Fine-tune, Backbone=ViT-L/142023.10 | 289.4 | |
| CLIPNoise Rate=20%, Evaluation Protocol=Fine-tune, Backbone=ViT-B/322023.10 | 236.3 |