Image Captioning on M3IT COCO
38.8Rouge-Lv-LCM
Evaluation Results
| Method | Links | |
|---|---|---|
| v-LCMvariant=v-LCM2026.03 | 38.8 | |
| Percep. LMModel Size=8B2026.03 | 36.3 | |
| Percep. LMModel Size=3B2026.03 | 34.3 | |
| Percep. LMModel Size=1B2026.03 | 27.5 | |
| Qwen2.5-VLModel Size=3B2026.03 | 25.1 | |
| Qwen2-VLModel Size=2B2026.03 | 24.9 | |
| Qwen2-VLModel Size=7B2026.03 | 23.7 | |
| InterVL2Model Size=8B2026.03 | 21 | |
| Qwen2.5-VLModel Size=7B2026.03 | 18.5 | |
| LCMvariant=LCM2026.03 | 18 | |
| InterVL2Model Size=4B2026.03 | 17.5 | |
| InternVL-2.5Model Size=8B2026.03 | 16.8 | |
| InternVL-2.5Model Size=4B2026.03 | 15.1 | |
| InternVL-2.5Model Size=1B2026.03 | 13.2 | |
| InterVL2Model Size=1B2026.03 | 12.6 |