Video Captioning on PE-Video
31.2ROUGE-LQwen2-VL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen2-VLModel Size=2B2026.03 | 31.2 | 37.3 | |
| Qwen2.5-VLModel Size=3B2026.03 | 28.9 | 34.4 | |
| Percep. LMModel Size=8B2026.03 | 27.4 | 31.9 | |
| v-LCMvariant=v-LCM2026.03 | 27.4 | 30 | |
| Qwen2-VLModel Size=7B2026.03 | 26.9 | 32.6 | |
| Percep. LMModel Size=1B2026.03 | 26.6 | 31 | |
| Percep. LMModel Size=3B2026.03 | 26.4 | 31.3 | |
| InterVL2Model Size=1B2026.03 | 25.5 | 31.5 | |
| LCMvariant=LCM2026.03 | 25.5 | 27.9 | |
| InternVL-2.5Model Size=1B2026.03 | 23.4 | 29.3 | |
| Qwen2.5-VLModel Size=7B2026.03 | 22.2 | 25.9 | |
| InternVL-2.5Model Size=8B2026.03 | 21.4 | 26 | |
| InterVL2Model Size=8B2026.03 | 18.6 | 23.4 | |
| InterVL2Model Size=4B2026.03 | 15 | 18.4 | |
| InternVL-2.5Model Size=4B2026.03 | 14.6 | 18 |