Video Recognition on HMDB-51 (Base/Novel/HM Metrics)
77.9Base AccuracyViCLIP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ViCLIPMode=Tuning, GFLOPs=1612024.11 | 77.9 | 52.4 | 62.7 | |
| Zero-ViCLIP + AdaptFormerMode=Efficient transfer, GFLOPs=179.802024.11 | 77.5 | 54.9 | 64.3 | |
| Zero-ViCLIP + MSTA + LccMode=Efficient transfer, GFLOPs=179.842024.11 | 77.5 | 57.9 | 66.3 | |
| Zero-ViCLIP + LORAMode=Efficient transfer, GFLOPs=179.802024.11 | 77.2 | 54.6 | 64 | |
| Zero-ViCLIP + MSTAMode=Efficient transfer, GFLOPs=179.842024.11 | 77.2 | 57.5 | 65.9 | |
| OSTMode=Tuning2024.11 | 75.1 | 34.4 | 47.2 | |
| ViFi-CLIPMode=Tuning, GFLOPs=2812024.11 | 74.5 | 53.9 | 62.5 | |
| Vanilla CLIPMode=Tuning, GFLOPs=2812024.11 | 71.5 | 55.3 | 62.4 | |
| XCLIPMode=Adapting, GFLOPs=1452024.11 | 69.9 | 45.3 | 55 | |
| ActionCLIPMode=Adapting, GFLOPs=2822024.11 | 69.3 | 38.3 | 49.3 | |
| Zero-ViCLIPMode=Efficient transfer, GFLOPs=161.742024.11 | 62.5 | 48.9 | 54.9 | |
| Vanilla CLIPMode=Zero-shot, GFLOPs=2812024.11 | 50.6 | 48.1 | 49.3 | |
| A5Mode=Adapting, GFLOPs=2842024.11 | 46.9 | 16.5 | 24.4 |