| object-categorization-on-grit | CLIP | #1 | Categorization (ablation): 48.1 |
| action-recognition-on-rareact | CLIP | #2 | mWAP: 40.7 |
| few-shot-image-classification-on-imagenet-0-shot | CLIP (ViT B/32) | #2 | Accuracy: 63.2% |
| few-shot-image-classification-on-imagenet-0-shot | CLIP (ResNet50) | #3 | Accuracy: 59.6% |
| long-tail-learning-on-coco-mlt | CLIP(ViT-B/16) | #2 | Average mAP: 60.17 |
| long-tail-learning-on-coco-mlt | CLIP(ResNet-50) | #4 | Average mAP: 56.19 |
| long-tail-learning-on-voc-mlt | CLIP(ViT-B/16) | #2 | Average mAP: 85.77 |
| long-tail-learning-on-voc-mlt | CLIP(ResNet-50) | #4 | Average mAP: 84.30 |
| meme-classification-on-multioff | CLIP | #3 | Accuracy: 62.4F1: 48.1 |
| semi-supervised-image-classification-on-imagenet-0-2-labeled-data | CLIP (ResNet-50) | #3 | ImageNet Top-1 Accuracy: 40% |
| text-based-person-retrieval-with-noisy-correspondence-on-cuhk-pedes | CLIP-C | #4 | Rank-1: 66.41Rank 10: 90.89Rank-5: 85.15mAP: 59.36mINP: 43.02 |
| text-based-person-retrieval-with-noisy-correspondence-on-icfg-pedes | CLIP-C | #4 | Rank 1: 55.25Rank-10: 81.32Rank-5: 74.76mAP: 31.09mINP: 4.94 |
| text-based-person-retrieval-with-noisy-correspondence-on-rstpreid | CLIP-C | #4 | Rank 1: 54.45Rank 10: 86.70Rank 5: 77.80mAP: 42.58mINP: 21.38 |
| hateful-meme-classification-on-harm-p | CLIP | #5 | Accuracy: 80.6F1: 80.3 |
| image-classification-on-omnibenchmark | CLIP-RN50 | #5 | Average Top-1 Accuracy: 42.1 |
| image-to-text-retrieval-on-coco | CLIP (zero-shot) | #5 | Recall@1: 58.4Recall@5: 81.5Recall@10: 88.1 |
| object-recognition-on-shape-bias | CLIP (ViT-B) | #6 | shape bias: 79.9 |
| zero-shot-transfer-image-classification-on-sun | CLIP | #6 | Accuracy: 58.5 |
| hateful-meme-classification-on-pridemm | CLIP (fine-tuned) | #7 | Accuracy: 72.4F1: 72.3 |
| open-vocabulary-attribute-detection-on-ovad-box | CLIP VIT-B16 | #7 | mean average precision: 16.6 |
| zero-shot-transfer-image-classification-on-objectnet | CLIP | #8 | Accuracy (Private): 72.3 |
| zero-shot-transfer-image-classification-on-imagenet-v2 | CLIP | #9 | Accuracy (Private): 70.1 |
| zero-shot-transfer-image-classification-on-imagenet-a | CLIP | #12 | Accuracy (Private): 77.2 |
| zero-shot-transfer-image-classification-on-imagenet-r | CLIP | #12 | Accuracy: 88.9 |
| zero-shot-transfer-image-classification-on-imagenet | CLIP（ViT-L/14-336px） | #13 | Accuracy (Private): 76.2 |
| zero-shot-transfer-image-classification-on-imagenet | CLIP (ResNet50) | #17 | Accuracy (Private): 59.6 |
| zero-shot-transfer-image-classification-on-imagenet | CLIP | #18 | Accuracy (Public): 31.3 |
| image-classification-on-objectnet | CLIP | #15 | Top-1 Accuracy: 72.3 |
| prompt-engineering-on-ucf101 | CLIP | #15 | Harmonic mean: 73.85 |
| zero-shot-cross-modal-retrieval-on-flickr30k | CLIP | #17 | Image-to-text R@1: 88.0Image-to-text R@5: 98.7… |
| meme-classification-on-hateful-memes | CLIP (zero-shot) | #18 | ROC-AUC: 0.661 |
| prompt-engineering-on-imagenet-v2 | CLIP | #18 | Top-1 accuracy %: 60.83 |
| prompt-engineering-on-imagenet-a | CLIP | #20 | Top-1 accuracy %: 47.77 |
| prompt-engineering-on-imagenet-r | CLIP | #21 | Top-1 accuracy %: 73.96 |
| prompt-engineering-on-imagenet-s | CLIP | #21 | Top-1 accuracy %: 46.15 |
| zero-shot-cross-modal-retrieval-on-coco-2014 | CLIP | #24 | Image-to-text R@1: 58.4Image-to-text R@5: 81.5… |
| prompt-engineering-on-caltech-101 | CLIP | #28 | Harmonic mean: 95.40 |
| prompt-engineering-on-eurosat | CLIP | #28 | Harmonic mean: 60.03 |
| prompt-engineering-on-fgvc-aircraft | CLIP | #28 | Harmonic mean: 31.09 |
| prompt-engineering-on-oxford-iiit-pet-dataset | CLIP | #28 | Harmonic mean: 94.12 |
| prompt-engineering-on-stanford-cars | CLIP | #28 | Harmonic mean: 68.65 |
| prompt-engineering-on-sun397 | CLIP | #28 | Harmonic mean: 72.23 |
| prompt-engineering-on-dtd | CLIP | #29 | Harmonic mean: 56.37 |
| prompt-engineering-on-imagenet | CLIP | #29 | Harmonic mean: 70.22 |
| prompt-engineering-on-oxford-102-flower | CLIP | #29 | Harmonic mean: 74.83 |