Time-aware Dense Captioning on OmniDCBench 1.0 (test)
12.4Camera ScoreTimeChat-Captioner-7B-GRPO
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| TimeChat-Captioner-7B-GRPOModality=V + A, Parameters=7B, Training Strategy=Group Relative Policy Optimization2026.02 | 12.4 | 39.6 | 49.2 | 38.2 | 43.5 | 54.3 | 35 | |
| Gemini-2.5-FlashModality=V + A2026.02 | 11.5 | 38.1 | 42.1 | 22.4 | 27.6 | 42.6 | 30 | |
| TimeChat-Captioner-7B-SFTModality=V + A, Parameters=7B, Training Strategy=Supervised Fine-Tuning2026.02 | 8.9 | 30.5 | 36.3 | 30.6 | 33.9 | 44.8 | 32.6 | |
| Gemini-2.5-ProModality=V + A2026.02 | 8.1 | 48.1 | 39.1 | 25.4 | 34.5 | 46.4 | 33.7 | |
| Qwen3-OmniModality=V + A, Parameters=30B-A3B2026.02 | 3.1 | 20.2 | 21.6 | 5.1 | 14.1 | 25.4 | 14.3 | |
| Qwen2.5-OmniModality=V + A, Parameters=7B2026.02 | 1.6 | 3.9 | 12.3 | 3.4 | 3.3 | 15.7 | 4.6 | |
| OmniVinciModality=V + A, Parameters=9B2026.02 | 1.6 | 8.2 | 15.7 | 1.3 | 6.7 | 14.3 | 6.9 | |
| MiniCPM-o-2.6Modality=V + A, Parameters=8B2026.02 | 1.2 | 5.3 | 16.5 | 1.5 | 7.4 | 11.3 | 5.4 | |
| LongVALEModality=V + A, Parameters=7B, Specialization=Temporal-aware expert2026.02 | 0.8 | 0.6 | 1.3 | 0.5 | 3.1 | 5 | 1.8 |