List your company

Visual Question Answering on ScienceQA SQA-IMG

74.4SQAIMG AccuracyDART
62.33665.46868.671.732May 20, 2026May 24, 2026May 29, 2026Jun 3, 2026Jun 7, 2026Jun 12, 2026Jun 17, 2026

Evaluation results

MethodLinks
DARTBackbone=LLaVA-1.5-13B, Visual Token Retention=128 tokens2026.0674.497.1
SparseVLMBackbone=LLaVA-1.5-13B, Visual Token Retention=128 tokens2026.0674.197.9
DARTBackbone=LLaVA-1.5-13B, Visual Token Retention=64 tokens2026.0673.992.5
SPAREBackbone=LLaVA-1.5-13B, Visual Token Retention=32 tokens2026.0673.294.7
LLaVA-1.5-13BBackbone=LLaVA-1.5-13B, Visual Token Retention=576 tokens2026.0672.8100
SPAREBackbone=LLaVA-1.5-13B, Visual Token Retention=128 tokens2026.0672.898
DARTBackbone=LLaVA-1.5-13B, Visual Token Retention=32 tokens2026.0672.886
CDPrunerBackbone=LLaVA-1.5-13B, Visual Token Retention=128 tokens2026.0672.697.9
SparseVLMBackbone=LLaVA-1.5-13B, Visual Token Retention=64 tokens2026.0672.589.7
CDPrunerBackbone=LLaVA-1.5-13B, Visual Token Retention=64 tokens2026.0672.596.3
SPAREBackbone=LLaVA-1.5-13B, Visual Token Retention=64 tokens2026.0672.397.2
CDPrunerBackbone=LLaVA-1.5-13B, Visual Token Retention=32 tokens2026.0671.992.6
PruMerge+Backbone=LLaVA-1.5-13B, Visual Token Retention=128 tokens2026.0671.893.2
PruMerge+Backbone=LLaVA-1.5-13B, Visual Token Retention=64 tokens2026.0671.790
PyramidDropModel=LLaVA-Next-7B, Token Budget=320, Reduction Ratio=88.9%2026.0569.590.1
DARTModel=LLaVA-Next-7B, Token Budget=320, Reduction Ratio=88.9%2026.0568.396.4
SPprunerModel=LLaVA-Next-7B, Token Budget=320, Reduction Ratio=88.9%2026.056897.7
DivPruneModel=LLaVA-Next-7B, Token Budget=320, Reduction Ratio=88.9%2026.0567.796.1
VanillaModel=LLaVA-Next-7B, Token Budget=2880, Reduction Ratio=0%2026.0567.6100
VisionZipModel=LLaVA-Next-7B, Token Budget=320, Reduction Ratio=88.9%2026.0567.696.2
FastVModel=LLaVA-Next-7B, Token Budget=320, Reduction Ratio=88.9%2026.0562.886.8