Video Recognition on Kinetics-600
74.5Top-1 AccViCLIP + MSTA + LCC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ViCLIP + MSTA + LCCFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 74.5 | — | |
| ViCLIP + MSTAFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 74.1 | — | |
| OSTFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 73.9 | — | |
| ViCLIP + LORAFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 73.8 | — | |
| ViCLIP + AdaptFormerFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 72.8 | — | |
| ViCLIPFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 71.5 | — | |
| ViFi-CLIPFrames=32, Zero-shot transfer=true, Inference=single-view2024.11 | 71.2 | — | |
| Vita-CLIPFrames=32, Zero-shot transfer=true, Inference=single-view2024.11 | 67.4 | — | |
| XCLIPFrames=32, Zero-shot transfer=true, Inference=single-view2024.11 | 65.2 | — | |
| Vanilla CLIPFrames=32, Zero-shot transfer=true, Inference=single-view2024.11 | 59.8 | — | |
| Zero-ViCLIPFrames=8, Zero-shot transfer=true, Inference=single-view2024.11 | 59.2 | — | |
| A5Frames=8/32, Zero-shot transfer=true, Inference=single-view2024.11 | 55.8 | — | |
| ER-ZSARFrames=16, Zero-shot transfer=true, Inference=single-view2024.11 | 42.1 | — |