List your company
ResearchFormerly WizWand

Multi-modal Understanding on SEED-Bench (overall)

62.9Overall ScoreCSR
31.49239.64647.855.954Nov 7, 2023Jan 31, 2024Apr 25, 2024Jul 19, 2024Oct 12, 2024Jan 5, 2025Mar 31, 2025
Updated 8mo ago

Evaluation results

MethodLinks
CSRBackbone=LLaVA-1.5-13B2024.0562.9
Self-rewardingBackbone=LLaVA-1.5-13B2024.0562.8
LLaVA-1.5-13BModel Size=13B2024.0561.6
CogVLM2025.0361.22
Task Arithmeticunsupervised=false2025.0360.85
CSRBackbone=LLaVA-1.5-7B2024.0560.3
POVIDBackbone=LLaVA-1.5-7B2024.0560.2
AdaMMSunsupervised=true2025.0360.16
RLHF-VBackbone=LLaVA-1.5-7B2024.0560.1
Self-rewardingBackbone=LLaVA-1.5-7B2024.0560
DARE-Linearunsupervised=false2025.0359.81
mPLUG-Owl2(base)2025.0359.41
VlfeedbackBackbone=LLaVA-1.5-7B2024.0559.3
LLaVA-1.5-HACLVision Encoder=VIT-L (0.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1258.9
LLaVA-1.5Vision Encoder=ViT-L (0.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1158.6
LLaVA-1.5Vision Encoder=VIT-L (0.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1258.6
LLaVA-1.5-7BModel Size=7B2024.0558.6
Qwen-VL-ChatVision Encoder=ViT-G (1.9B), Language Model=Qwen (7B), Zero-shot=true2023.1158.2
Qwen-VL-ChatVision Encoder=ViT-G (1.9B), Language Model=Qwen (7B), Zero-shot=true2023.1258.2
Human-PreferBackbone=LLaVA-1.5-7B2024.0558.1
mPLUG-Owl2Vision Encoder=ViT-L (0.3B), Language Model=LLAMA (7B), Zero-shot=true2023.1157.8
DARE-Tiesunsupervised=false2025.0357.62
InstructBLIPVision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1153.4
InstructBLIPVision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1253.4
Ties-Mergingunsupervised=false2025.0352.32
MetaGPTunsupervised=true2025.0350.81
BLIP-2Vision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1146.4
BLIP-2Vision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1246.4
MiniGPT-4Vision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1142.8
MiniGPT-4Vision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1242.8
MiniGPT-4-HACLVision Encoder=ViT-g (1.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1242.5
mPLUG-OwlVision Encoder=ViT-L (0.3B), Language Model=LLaMA (7B), Zero-shot=true2023.1134
mPLUG-OwlVision Encoder=ViT-L (0.3B), Language Model=LLaMA (7B), Zero-shot=true2023.1234
LLaVA-HACLVision Encoder=ViT-L (0.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1233.9
LLaVAVision Encoder=ViT-L (0.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1133.5
LLaVAVision Encoder=VIT-L (0.3B), Language Model=Vicuna (7B), Zero-shot=true2023.1233.5
OtterVision Encoder=ViT-L (0.3B), Language Model=LLaMA (7B), Zero-shot=true2023.1132.9
OtterVision Encoder=VIT-L (0.3B), Language Model=LLaMA (7B), Zero-shot=true2023.1232.9
LLaMA-Adapter-v2Vision Encoder=ViT-L (0.3B), Language Model=LLaMA (7B), Zero-shot=true2023.1132.7
LLaMA-Adapter-v2Vision Encoder=ViT-L (0.3B), Language Model=LLaMA (7B), Zero-shot=true2023.1232.7