List your company
ResearchFormerly WizWand

Meme Classification on MultiOFF

71.1AccuracyQwen2-VL-7B
37.61246.3065563.694Feb 18, 2025
Updated 8mo ago

Evaluation results

MethodLinks
Qwen2-VL-7BEvaluation Protocol=RA-HMD2025.0271.164.8
LLaVA-1.5-7BEvaluation Protocol=RA-HMD2025.0270.963.6
Qwen2-VL-2BEvaluation Protocol=RA-HMD2025.0268.561.8
LLaVA-1.5-7BEvaluation Protocol=SFT2025.0267.857.8
Qwen2-VL-7BEvaluation Protocol=SFT2025.0267.855.5
RGCL2025.0267.158.1
Qwen2-VL-2BEvaluation Protocol=4-shot2025.0267.144.9
Qwen2-VL-2BEvaluation Protocol=SFT2025.0266.454.5
RA-HMDbackbone=Qwen2-VL-2B, fine-tuning set=HatefulMemes, protocol=RKC2025.026663.4
Qwen2-VL-7BEvaluation Protocol=4-shot2025.0264.454.7
RA-HMDbackbone=Qwen2-VL-7B, fine-tuning set=HatefulMemes, protocol=RKC2025.0263.855.6
Qwen2-VL-7BEvaluation Protocol=zero-shot2025.0263.435.9
Qwen2-VL-7Bfine-tuning set=HatefulMemes, protocol=SFT + zero-shot2025.0263.129.7
Qwen2-VL-2BEvaluation Protocol=zero-shot2025.0263.136.3
LLaVA-1.5-7Bfine-tuning set=HatefulMemes, protocol=SFT + zero-shot2025.0262.832.5
RA-HMDbackbone=LLaVA-1.5-7B, fine-tuning set=HatefulMemes, protocol=RKC2025.0262.451.7
CLIPTraining=Supervised fine-tuned2025.0262.448.1
HateCLIPperTraining=Supervised fine-tuned2025.0262.454.8
Qwen2-VL-2Bfine-tuning set=HatefulMemes, protocol=SFT + zero-shot2025.0262.329.3
Qwen2-VL-7Bfine-tuning set=HatefulMemes, protocol=SFT + few-shot2025.026253.7
GPT-4omode=low-resourced2025.0261.151.1
LLaVA-1.5-7BEvaluation Protocol=zero-shot2025.0259.651.7
GPT-4o2025.0258.358.1
LLaVA-1.5-7Bfine-tuning set=HatefulMemes, protocol=SFT + few-shot2025.025638.9
RGCLfine-tuning set=HatefulMemes2025.0253.745.1
Qwen2-VL-2Bfine-tuning set=HatefulMemes, protocol=SFT + few-shot2025.0253.742.4
LLaVA-1.5-7BEvaluation Protocol=4-shot2025.0238.956