Video Recognition on Something-Something V2
22.2Base ScoreZero-ViCLIP + MSTA + Lcc
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Zero-ViCLIP + MSTA + LccMode=Efficient transfer, GFLOPs=179.842024.11 | 22.2 | 16.5 | 18.9 | |
| Zero-ViCLIP + MSTAMode=Efficient transfer, GFLOPs=179.842024.11 | 21.9 | 16.3 | 18.7 | |
| Zero-ViCLIP + AdaptFormerMode=Efficient transfer, GFLOPs=179.802024.11 | 20.4 | 14.6 | 17 | |
| ViCLIPMode=Tuning, GFLOPs=1612024.11 | 19.4 | 15.2 | 17 | |
| Zero-ViCLIP + LORAMode=Efficient transfer, GFLOPs=179.802024.11 | 18.5 | 13.6 | 16 | |
| FROSTEREvaluation Protocol=Base-to-novel2026.06 | 18.3 | 12.2 | 14.6 | |
| TACOEvaluation Protocol=Base-to-novel2026.06 | 17.7 | 12.4 | 14.6 | |
| Open-MeDeEvaluation Protocol=Base-to-novel2026.06 | 17.1 | 12.3 | 14.3 | |
| ViFi-CLIPMode=Tuning, GFLOPs=2812024.11 | 16.5 | 12.4 | 14.2 | |
| ViFi-CLIPEvaluation Protocol=Base-to-novel2026.06 | 16.2 | 12.1 | 13.9 | |
| Open-VCLIPEvaluation Protocol=Base-to-novel2026.06 | 16 | 11 | 13 | |
| OSTMode=Tuning2024.11 | 15.3 | 11.5 | 13.1 | |
| ActionCLIPMode=Adapting, GFLOPs=2822024.11 | 13.5 | 10.4 | 11.7 | |
| ActionCLIPEvaluation Protocol=Base-to-novel2026.06 | 13.3 | 10.1 | 11.5 | |
| Vanilla CLIPMode=Tuning, GFLOPs=2812024.11 | 13.2 | 10.8 | 11.9 | |
| Zero-ViCLIPMode=Efficient transfer, GFLOPs=161.742024.11 | 9.3 | 8.9 | 9.1 | |
| ST-AdapterEvaluation Protocol=Base-to-novel2026.06 | 9.3 | 8.4 | 8.8 | |
| XCLIPMode=Adapting, GFLOPs=1452024.11 | 8.9 | 6.8 | 7.7 | |
| A5Mode=Adapting, GFLOPs=2842024.11 | 8.6 | 6.4 | 7.3 | |
| X-CLIPEvaluation Protocol=Base-to-novel2026.06 | 8.5 | 6.6 | 7.4 | |
| AIMEvaluation Protocol=Base-to-novel2026.06 | 8.5 | 7.9 | 8.2 | |
| VPTEvaluation Protocol=Base-to-novel2026.06 | 8.3 | 5.3 | 6.4 | |
| CLIPEvaluation Protocol=Base-to-novel2026.06 | 4.9 | 5.3 | 5.1 | |
| Vanilla CLIPMode=Zero-shot, GFLOPs=2812024.11 | 4.8 | 5.4 | 5.1 |