ActivityNet-QA
ActivityNet-QA: Evaluates temporal, video, speech, or audio understanding beyond static text and image inputs.
3rows
overall_accuracyprimary metric
2026-05-27sampled
Metadata
Metrics
Overall accuracy, Motion accuracy, Spatial relation accuracy, Temporal relation accuracy, Free-form accuracy, WUPS@0.9, WUPS@0.0
| Rank | Subject | Overall accuracy | Model Match | Provenance | Sampled |
|---|---|---|---|---|---|
| 1 | E-SA | 31.8% | — | Imported | 2026-05-27 |
| 2 | E-MN | 27.1% | — | Imported | 2026-05-27 |
| 3 | E-VQA | 25.1% | — | Imported | 2026-05-27 |
No matching rows.