List your company

Visual Question Answering on ScienceQA image

75.1AccuracyCSR
59.91663.85867.871.742May 23, 2024Sep 28, 2024Feb 3, 2025Jun 12, 2025Oct 18, 2025Feb 23, 2026Jul 2, 2026

Evaluation results

MethodLinks
CSRBackbone=LLaVA-1.5-13B2024.0575.1
Self-rewardingBackbone=LLaVA-1.5-13B2024.0574.3
SparseVLM#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0374.3
FastV#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0374.2
DART#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0374.2
ToFuLLM=Vicuna-13B, compression_ratio_r=0.9252024.0574
PITOMELLM=Vicuna-13B, compression_ratio_r=0.9252024.0574
ToMeLLM=Vicuna-13B, compression_ratio_r=0.9252024.0573.8
VisionZip#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0373.8
MQT-LLaVA#Vision Tokens=144, Backbone=LLaVA-1.5-13B2026.0373.6
PyramidDrop#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0373.3
Prumerge+#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0373.3
CDPruner#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0373.2
M3#Vision Tokens=144, Backbone=LLaVA-1.5-13B2026.0372.9
LLaVA-1.5-13BLLM=Vicuna-13B2024.0572.8
LLaVA-1.5-13B#Vision Tokens=576, Backbone=LLaVA-1.5-13B2026.0372.8
DivPrune#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0372.8
VisPruner#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0372.8
TRIM#Vision Tokens=128, Backbone=LLaVA-1.5-13B2026.0372.4
FMVR-LLaVA#Vision Tokens=576, Backbone=LLaVA-1.5-13B2026.0372.4
FMVR-LLaVA#Vision Tokens=144, Backbone=LLaVA-1.5-13B2026.0372
LLaVA-1.5-13BModel Size=13B2024.0571.6
FMVR-LLaVA#Vision Tokens=36, Backbone=LLaVA-1.5-13B2026.0371.5
FMVR-LLaVA#Vision Tokens=9, Backbone=LLaVA-1.5-13B2026.0370.9
CSRBackbone=LLaVA-1.5-7B2024.0570.7
DiffRateLLM=Vicuna-13B, compression_ratio_r=0.9252024.0570.6
LLaVA-1.5-7BLLM=Vicuna-7B2024.0570.4
DCTLLM=Vicuna-13B, compression_ratio_r=0.9252024.0570.3
Self-rewardingBackbone=LLaVA-1.5-7B2024.0569.6
LLaVA-1.5-7BRetained Tokens=576, Retention Ratio=100%, Base Model=LLaVA-1.5-7B2025.0869.5
MonkeyLLM=Qwen-7B2024.0869.4
FMVR-LLaVA#Vision Tokens=1, Backbone=LLaVA-1.5-13B2026.0369.4
PITOMELLM=Vicuna-7B, compression_ratio_r=0.92024.0569
FASTLLM=Vicuna-7B2024.0868.9
HiPruneRetained Tokens=192, Retention Ratio=33.3%, Base Model=LLaVA-1.5-7B2025.0868.9
HiPruneRetained Tokens=64, Retention Ratio=11.1%, Base Model=LLaVA-1.5-7B2025.0868.9
HiPrune++Retained Tokens=64, Retention Ratio=11.1%, Base Model=LLaVA-1.5-7B2025.0868.9
POVIDBackbone=LLaVA-1.5-7B2024.0568.8
HiPrune++Retained Tokens=192, Retention Ratio=33.3%, Base Model=LLaVA-1.5-7B2025.0868.8
ToMeLLM=Vicuna-7B, compression_ratio_r=0.92024.0568.7
mPLUG-Owl2LLM=LLaMA-7B2024.0868.7
HiPruneRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0768.7
ToFuLLM=Vicuna-7B, compression_ratio_r=0.92024.0568.5
HiPrune++Retained Tokens=128, Retention Ratio=22.2%, Base Model=LLaVA-1.5-7B2025.0868.5
HiPruneRetained Tokens=128, Retention Ratio=22.2%, Base Model=LLaVA-1.5-7B2025.0868.3
Qwen-VL-ChatLLM=Qwen-7B2024.0868.2
Chain of SpotLLM=Vicuna-7B2024.0868.2
DARTRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0768
HiPruneRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0768
EADPRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0768
CDPrunerRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.9
DivPruneRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.8
CDPrunerRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.7
VisionZipRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0767.7
LLaVA-NeXT-7BRetained Tokens=2,880, Token Reduction Ratio=100%2026.0767.6
EADPRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.6
VisionZipRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.5
PruMerge+Retained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.5
DivPruneRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.5
DARTRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0767.5
CDPrunerRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0767.5
PruMerge+Retained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.4
EADPRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0767.4
SparseVLMRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.3
VisionZipRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.3
HiPruneRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.3
FastVRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.2
RLHF-VBackbone=LLaVA-1.5-7B2024.0567.1
TRIMRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0767.1
DARTRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767.1
DivPruneRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0767.1
SparseVLMRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0767
PDropRetained Tokens=640, Token Reduction Ratio=↓ 77.8%2026.0766.9
LLaVA-1.5-7BModel Size=7B2024.0566.8
LLaVA-v1.5LLM=Vicuna-7B2024.0866.8
PruMerge+Retained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0766.7
PDropRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0766.5
DiffRateLLM=Vicuna-7B, compression_ratio_r=0.92024.0566.4
TRIMRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0766.4
FastVRetained Tokens=320, Token Reduction Ratio=↓ 88.9%2026.0766.3
VlfeedbackBackbone=LLaVA-1.5-7B2024.0566.2
DCTLLM=Vicuna-7B, compression_ratio_r=0.92024.0565.8
Human-PreferBackbone=LLaVA-1.5-7B2024.0565.8
TRIMRetained Tokens=160, Token Reduction Ratio=↓ 94.4%2026.0765.7
InstructBLIPLLM=Vicuna-13B2024.0563.1
InstructBLIPLLM=Vicuna-13B2024.0863.1
BLIP-2LLM=Vicuna-7B2024.0561
BLIP-2LLM=Vicuna-13B2024.0861
InstructBLIPLLM=Vicuna-7B2024.0560.5