Video Action Recognition on HMDB-51 (val)
62.4Top-1 Acc (K=2)EZ-CLIP
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| EZ-CLIPBackbone=ViT-142023.12 | 62.4 | 65.3 | 68.4 | 70.7 | |
| EZ-CLIPBackbone=ViT-162023.12 | 57.3 | 61.1 | 65.4 | 67.7 | |
| ViFi-CLIP2023.12 | 57.2 | 62.7 | 64.5 | 66.8 | |
| X-CLIP2023.12 | 53 | 57.3 | 62.8 | 62.4 | |
| EZ-CLIPBackbone=ViT-322023.12 | 52.4 | 56.9 | 60.4 | 63.7 | |
| ActionCLIP2023.12 | 47.5 | 57.3 | 57.9 | 59.1 | |
| Vanilla CLIP2023.12 | 41.9 | 41.9 | 41.9 | 41.9 | |
| A52023.12 | 39.7 | 50.7 | 57 | 62.4 |