Image Understanding Evaluation on MME
1,862MME ScoreVanilla
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VanillaToken Budget (Retained Tokens)=576, Reduction Ratio (↓ %)=0%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,862 | 100 | |
| SPprunerToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,849 | 98.5 | |
| DARTToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,844 | 97.2 | |
| SPprunerToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,841 | 99.4 | |
| DARTToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,834 | 99 | |
| PyramidDropToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,791 | 98.7 | |
| PACTToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,790 | 98.1 | |
| VisionZipToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,769 | 99 | |
| DivPruneToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,765 | 98.8 | |
| PyramidDropToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,765 | 96.5 | |
| VisionZipToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,758 | 97.6 | |
| PACTToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,757 | 96.9 | |
| SPprunerToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,750 | 95.2 | |
| DARTToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,749 | 93.4 | |
| DivPruneToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,711 | 97.4 | |
| VisionZipToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,688 | 90.4 | |
| DivPruneToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,615 | 94.4 | |
| FastVToken Budget (Retained Tokens)=192, Reduction Ratio (↓ %)=66.7%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,612 | 90 | |
| PyramidDropToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,579 | 88.6 | |
| PACTToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,564 | 94 | |
| FastVToken Budget (Retained Tokens)=128, Reduction Ratio (↓ %)=77.8%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,490 | 82.9 | |
| FastVToken Budget (Retained Tokens)=64, Reduction Ratio (↓ %)=88.9%, Base Model (LLaVA-1.5-7B)=LLaVA-1.5-7B2026.05 | 1,256 | 74 |