Dense Video Captioning on ChronusAV
63.2F1 ScorePadCaptioner
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PadCaptionerNumber of Parameters=3B2026.07 | 63.2 | 40 | 12.4 | 9.6 | 12.4 | |
| PadCaptioner (3B)T/video=7526.7, T/token=13.72026.07 | 63.2 | 40 | — | — | — | |
| PadCaptioner-Number of Parameters=3B, Training Budget=12K videos2026.07 | 61.8 | 38.8 | 10.5 | 7.8 | 10.6 | |
| ChronusOmniNumber of Parameters=7B2026.07 | 60.1 | 36.8 | 8.6 | 2.9 | 7.6 | |
| ChronusOmni (7B)T/video=28093.0, T/token=41.02026.07 | 60.1 | 36.8 | — | — | — | |
| Qwen3-OmniNumber of Parameters=30B-A3B2026.07 | 51.8 | 37 | 5.4 | 2.5 | 6.5 | |
| video-SALMONN 2+Number of Parameters=3B2026.07 | 21.5 | 17.6 | 2.8 | 1.5 | 4.3 | |
| video-SALMONN-2+ (3B)T/video=3083.4, T/token=22.82026.07 | 21.5 | 17.6 | — | — | — | |
| LongVALE-LLMNumber of Parameters=7B2026.07 | 18.4 | 25.4 | 3.1 | 4.6 | 8.8 |