Video Temporal Grounding on ActivityNet
40.33mIoUQwen3-VL-4B
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Qwen3-VL-4BTokens (#)=10460, Tokens (%)=1002026.03 | 40.33 | 100 | 59.4 | 100 | — | 1,263.4 | 1,263.4 | 1 | |
| SemVIDTokens (#)=1307, Tokens (%)=12.52026.03 | 38.49 | 95.4 | 4.8 | 8.7 | 33.1 | 184.6 | 217.7 | 5.8 | |
| FastVIDTokens (#)=1370, Tokens (%)=13.12026.03 | 33.16 | 82.2 | 5.4 | 9.1 | 23.8 | 185.9 | 209.7 | 6 | |
| VanillaRet.=100%2026.05 | 30.7 | — | — | — | — | — | — | — | |
| OTT-VidRet.=25%2026.05 | 25.5 | — | — | — | — | — | — | — | |
| UniCompRet.=25%2026.05 | 24.5 | — | — | — | — | — | — | — | |
| HoliTomRet.=25%2026.05 | 24.3 | — | — | — | — | — | — | — | |
| OTT-VidRet.=20%2026.05 | 24 | — | — | — | — | — | — | — | |
| UniCompRet.=20%2026.05 | 22.7 | — | — | — | — | — | — | — | |
| HoliTomRet.=20%2026.05 | 22.3 | — | — | — | — | — | — | — | |
| OTT-VidRet.=15%2026.05 | 21.8 | — | — | — | — | — | — | — | |
| FastVIDRet.=25%2026.05 | 20.8 | — | — | — | — | — | — | — | |
| UniCompRet.=15%2026.05 | 20.2 | — | — | — | — | — | — | — | |
| FastVIDRet.=20%2026.05 | 19.9 | — | — | — | — | — | — | — | |
| VisionZipTokens (#)=1307, Tokens (%)=12.52026.03 | 19.89 | 49.3 | 4.8 | 8.7 | 710.9 | 184.3 | 895.2 | 1.4 | |
| HoliTomRet.=15%2026.05 | 19.5 | — | — | — | — | — | — | — | |
| OTT-VidRet.=10%2026.05 | 18.4 | — | — | — | — | — | — | — | |
| FastVIDRet.=15%2026.05 | 18.1 | — | — | — | — | — | — | — | |
| FlashVIDRet.=25%2026.05 | 16.6 | — | — | — | — | — | — | — | |
| UniCompRet.=10%2026.05 | 16 | — | — | — | — | — | — | — | |
| FastVIDRet.=10%2026.05 | 15.7 | — | — | — | — | — | — | — | |
| HoliTomRet.=10%2026.05 | 14.8 | — | — | — | — | — | — | — | |
| FlashVIDRet.=20%2026.05 | 14.2 | — | — | — | — | — | — | — | |
| FlashVIDRet.=15%2026.05 | 11.5 | — | — | — | — | — | — | — | |
| FlashVIDRet.=10%2026.05 | 9.8 | — | — | — | — | — | — | — |