Visual Question Answering on GQA (GQA and RelAcc.)
61.9GQA ScoreVanilla
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VanillaRetained Tokens=576, Base Model=LLaVA-v1.5-7B2026.04 | 61.9 | 100 | |
| TwigVLMRetained Tokens=128, Params.=610M, Base Model=LLaVA-v1.5-7B2026.04 | 60.6 | 99 | |
| LearnPrunerRetained Tokens=128, Params.=0.53M, Base Model=LLaVA-v1.5-7B2026.04 | 60.3 | 98.5 | |
| DivPruneRetained Tokens=128, Base Model=LLaVA-v1.5-7B2026.04 | 59.3 | 96.4 | |
| VisionZip ‡Retained Tokens=128, Params.=20.9M, Base Model=LLaVA-v1.5-7B2026.04 | 58.9 | 97.3 | |
| LearnPrunerRetained Tokens=64, Params.=0.53M, Base Model=LLaVA-v1.5-7B2026.04 | 58.9 | 96.9 | |
| TwigVLMRetained Tokens=64, Params.=610M, Base Model=LLaVA-v1.5-7B2026.04 | 58.8 | 96 | |
| VisPrunerRetained Tokens=128, Base Model=LLaVA-v1.5-7B2026.04 | 58.2 | 97.3 | |
| DARTRetained Tokens=128, Base Model=LLaVA-v1.5-7B2026.04 | 57.9 | 95.4 | |
| DivPruneRetained Tokens=64, Base Model=LLaVA-v1.5-7B2026.04 | 57.8 | 93.9 | |
| VisionZipRetained Tokens=128, Base Model=LLaVA-v1.5-7B2026.04 | 57.6 | 96.3 | |
| LearnPrunerRetained Tokens=32, Params.=0.53M, Base Model=LLaVA-v1.5-7B2026.04 | 57.2 | 94.8 | |
| VisionZip ‡Retained Tokens=64, Params.=20.9M, Base Model=LLaVA-v1.5-7B2026.04 | 57 | 95.1 | |
| SparseVLMRetained Tokens=128, Base Model=LLaVA-v1.5-7B2026.04 | 56 | 92.6 | |
| VisPrunerRetained Tokens=64, Base Model=LLaVA-v1.5-7B2026.04 | 55.4 | 94.3 | |
| VisionZipRetained Tokens=64, Base Model=LLaVA-v1.5-7B2026.04 | 55.1 | 93 | |
| DivPruneRetained Tokens=32, Base Model=LLaVA-v1.5-7B2026.04 | 54.9 | 90.5 | |
| DARTRetained Tokens=64, Base Model=LLaVA-v1.5-7B2026.04 | 54.7 | 91.9 | |
| DARTRetained Tokens=32, Base Model=LLaVA-v1.5-7B2026.04 | 52.9 | 88 | |
| SparseVLMRetained Tokens=64, Base Model=LLaVA-v1.5-7B2026.04 | 52.7 | 85.6 | |
| VisPrunerRetained Tokens=32, Base Model=LLaVA-v1.5-7B2026.04 | 52.2 | 89.7 | |
| VisionZipRetained Tokens=32, Base Model=LLaVA-v1.5-7B2026.04 | 51.8 | 87.8 | |
| FastVRetained Tokens=128, Base Model=LLaVA-v1.5-7B2026.04 | 49.6 | 82.1 | |
| SparseVLMRetained Tokens=32, Base Model=LLaVA-v1.5-7B2026.04 | 48.3 | 77.5 | |
| FastVRetained Tokens=64, Base Model=LLaVA-v1.5-7B2026.04 | 46.1 | 71.4 | |
| FastVRetained Tokens=32, Base Model=LLaVA-v1.5-7B2026.04 | 41.5 | 58.6 |