| action-classification-on-kinetics-400 | PolyViT-L/16x2 | – | Acc@1: 82.4 |
| action-classification-on-kinetics-600 | PolyViT-L/16x2 | – | Top-1 Accuracy: 82.9 |
| action-classification-on-moments-in-time | PolyViT-L/16x2 | – | Top 1 Accuracy: 38.6 |
| audio-classification-on-audioset | PolyViT-B | – | Test mAP: 0.445 |
| audio-classification-on-vggsound | PolyViT-B | – | Top 1 Accuracy: 55.1 |
| image-classification-on-cifar-10 | PolyViT-L | – | Percentage correct: 99.4 |
| image-classification-on-cifar-100 | PolyViT-L | – | Percentage correct: 93.9 |
| image-classification-on-imagenet | PolyViT-L | – | Top 1 Accuracy: 85.1% |
| linear-probe-image-classification-on-caltech-101 | PolyViT-Base (L_adapt=0) linear probe | – | Accuracy: 91.0 |
| linear-probe-image-classification-on-dtd | PolyViT-Base (image tasks) linear probe | – | Accuracy: 72.5 |
| linear-probe-image-classification-on-sun397 | PolyViT-Base (image tasks) linear probe | – | Accuracy: 80.0 |