Dense Video Captioning on ActivityNet Captions (test)
680METEORVTimeLLM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VTimeLLMModel size=7B2023.11 | 680 | — | 276 | 580 | |
| VTimeLLMModel size=13B2023.11 | 670 | — | 272 | 590 | |
| VideoChatGPTModel size=7B2023.11 | 210 | — | 58 | 190 | |
| VideoLLaMAModel size=7B2023.11 | 190 | — | 58 | 190 | |
| VideoChatModel size=7B2023.11 | 90 | — | 22 | 90 | |
| ESG moduleProposal source=Generated events, Training set=Enlarged2020.06 | 9.894 | — | — | — | |
| ESG moduleProposal source=Generated events, Training set=Official2020.06 | 9.336 | — | — | — | |
| MCCLPre-trained=No2026.03 | 9.05 | 2.68 | 34.92 | 6.16 | |
| DIBSPre-trained=Yes2026.03 | 8.93 | — | 31.89 | 5.85 | |
| JEDDi-Nettraining_strategy=joint training, context=with context2018.02 | 8.81 | — | — | — | |
| DDVCPre-trained=No2026.03 | 8.62 | 2.44 | 35.48 | 6.55 | |
| E2DVCPre-trained=No2026.03 | 8.57 | 2.43 | 33.63 | 6.13 | |
| CM2Pre-trained=No2026.03 | 8.55 | 2.38 | 33.01 | 6.18 | |
| Vid2SeqPre-trained=Yes2026.03 | 8.5 | — | 30.1 | 5.8 | |
| ROS-DVCPre-trained=No2026.03 | 8.45 | 2.36 | 35.04 | 6.45 | |
| PDVCPre-trained=No2026.03 | 8.06 | 2.21 | 29.97 | 5.92 |