Online Zero-Shot Temporal Action Localization on ActivityNet 1.3 (75%-25% class split)
11.63mAP@0.5VFEAL
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VFEALVisual Encoder=ViCLIP2026.05 | 11.63 | 3.4 | 34 | 5.13 | |
| Baseline-IIVisual Encoder=ViCLIP2026.05 | 1.16 | 0.35 | 12 | 0.54 | |
| Baseline-IVisual Encoder=CLIP2026.05 | 0.89 | 0.28 | 9 | 0.42 |