Few-Shot Action Recognition on UCF
97.5Accuracy (5-way 1-shot)TAMT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TAMTModality=Unimodal, Pre-training=ViT-B, Tuning=PEFT2024.11 | 97.5 | 98.8 | |
| CLIP-FSARModality=Multi-modal, Pre-training=CLIP-ViT-B, Tuning=FFT2024.11 | 96.6 | 99 | |
| OTAMModality=Unimodal, Pre-training=CLIP-ViT-B(V), Tuning=FFT2024.11 | 95.8 | 98.8 | |
| CLIP-CPM2CModality=Multi-modal, Pre-training=CLIP-ViT-B, Tuning=FFT2024.11 | 95 | 98.6 | |
| CapFSARModality=Multi-modal, Pre-training=BLIP-ViT-B, Tuning=FFT2024.11 | 93.1 | 97.7 | |
| HyRSMModality=Unimodal, Pre-training=BLIP-ViT-B(V), Tuning=FFT2024.11 | 91.6 | 96.9 | |
| MASTAFModality=Unimodal, Pre-training=JFT-ViT-B, Tuning=FFT2024.11 | 91.6 | — | |
| TRXModality=Unimodal, Pre-training=BLIP-ViT-B(V), Tuning=FFT2024.11 | 90.9 | 97.4 | |
| CLIPModality=Multi-modal, Pre-training=CLIP-ViT-B, Tuning=Frozen2024.11 | 89.7 | 95.7 |