Video Captioning / Summarization on VideoXum
29.1Rouge-LInterVL2
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InterVL2Model Size=8B2026.03 | 29.1 | 26.1 | |
| Percep. LMModel Size=3B2026.03 | 27 | 36.4 | |
| Percep. LMModel Size=8B2026.03 | 26.2 | 33.7 | |
| Qwen2-VLModel Size=7B2026.03 | 26 | 32.4 | |
| Qwen2.5-VLModel Size=3B2026.03 | 26 | 32.9 | |
| InternVL-2.5Model Size=8B2026.03 | 24.9 | 20.5 | |
| Qwen2.5-VLModel Size=7B2026.03 | 24.1 | 28.9 | |
| Qwen2-VLModel Size=2B2026.03 | 23.6 | 29.8 | |
| Percep. LMModel Size=1B2026.03 | 21.8 | 33.2 | |
| LCMvariant=LCM2026.03 | 21.5 | 22.1 | |
| v-LCMvariant=v-LCM2026.03 | 20.6 | 25.3 | |
| InternVL-2.5Model Size=4B2026.03 | 18.1 | 23 | |
| InternVL-2.5Model Size=1B2026.03 | 17.1 | 23.2 | |
| InterVL2Model Size=4B2026.03 | 15.6 | 17.4 | |
| InterVL2Model Size=1B2026.03 | 15.3 | 17.7 |