Long-video understanding on HourVideo (test)
0.021Inference Time (s)PSFR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PSFR2026.03 | 0.021 | — | |
| MaxInfo2026.03 | 0.038 | — | |
| Scene-Caption LLM Selector-sNumber of beams=12026.03 | 0.6 | 1 | |
| ViaRLNumber of beams=12026.03 | 2.6 | 16 | |
| Scene-Caption LLM Selector-mNumber of beams=52026.03 | 2.8 | 1 | |
| ReMEmbR2026.03 | 80 | — | |
| GenSNumber of beams=12026.03 | 103 | 16 |