Quan Sun, Yuxin Fang, Ledell Wu, Xinlong Wang, Yue Cao
Contrastive language-image pre-training, CLIP for short, has gained increasing attention for its potential in various scenarios. In this paper, we propose EVA-CLIP, a series of models that significantly improve the efficiency and effectiveness of CLIP training. Our approach incorporates new techniques for representation learning, optimization, and augmentation, enabling EVA-CLIP to achieve superior performance compared to previous CLIP models with the same number of parameters but significantly smaller training costs. Notably, our largest 5.0B-parameter EVA-02-CLIP-E/14+ with only 9 billion seen samples achieves 82.0 zero-shot top-1 accuracy on ImageNet-1K val. A smaller EVA-02-CLIP-L/14+ with only 430 million parameters and 6 billion seen samples achieves 80.4 zero-shot top-1 accuracy on ImageNet-1K val. To facilitate open access and open research, we release the complete suite of EVA-CLIP to the community at https://github.com/baaivision/EVA/tree/master/EVA-CLIP.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Image Classification | ObjectNet | Top-1 Accuracy | 79.6 | EVA-02-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | ImageNet V2 | Accuracy (Private) | 75.7 | EVA-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | ImageNet-A | Accuracy (Private) | 82.1 | EVA-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | ImageNet | Accuracy (Private) | 82 | EVA-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | ImageNet-R | Accuracy | 94.5 | EVA-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | Food-101 | Top 1 Accuracy | 94.9 | EVA-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | ObjectNet | Accuracy (Private) | 79.6 | EVA-CLIP-E/14+ |
| Zero-Shot Transfer Image Classification | ImageNet-Sketch | Accuracy (Private) | 71.6 | EVA-CLIP-E/14+ |
| Zero-Shot Action Recognition | UCF101 | Top-1 Accuracy | 83.1 | EVA-CLIP-E/14+ |