Visual Counting on Emoji-Count (test)
16.16MAELVLM-Count
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LVLM-CountBase model=Gemma 3 27B2024.12 | 16.16 | — | 21.27 | |
| LVLM-CountBase model=GPT-4o2024.12 | 16.57 | — | 33.11 | |
| Gemma 3 27B2024.12 | 21.39 | — | 24.04 | |
| GPT-4o2024.12 | 23.57 | — | 36.97 | |
| LVLM-CountBase model=Qwen2 VL 72B AWQ2024.12 | 24.43 | — | 43.38 | |
| TFOC2024.12 | 64.64 | — | 87.45 | |
| Qwen2 VL 72B AWQ2024.12 | 78.05 | — | 159.22 | |
| Number of the target segmentation masks2024.12 | 82.47 | — | 107.98 | |
| GPT-4o counting target segmentaion masks2024.12 | 107.72 | — | 162.12 | |
| CountGD2024.12 | 137.93 | — | 156.8 | |
| GroundingREC2024.12 | 143.22 | — | 158.74 | |
| DAVEprm2024.12 | 198.99 | — | 208.08 |