Home News Latest Papers Tutorials Datasets Wiki SOTA LLM Models GPU Leaderboard Events

English

Meme Classification On Hateful Memes

Metrics

Accuracy

ROC-AUC

Results

Performance results of various models on this benchmark

Model Name	Accuracy	ROC-AUC	Paper Title	Repository
LMM-RGCL (Qwen2-VL-7B)	0.821	0.911	Improved Fine-Tuning of Large Multimodal Models for Hateful Meme Detection
Flamingo (few-shot:32)	-	0.700	Flamingo: a Visual Language Model for Few-Shot Learning
Vilio	0.695	0.825	Vilio: State-of-the-art Visio-Linguistic Models applied to Hateful Memes
CLIP (zero-shot)	-	0.661	Learning Transferable Visual Models From Natural Language Supervision
Human	0.847	0.8265	The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes
SEER (RegNet10B)	-	0.734	Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision
Ron Zhu	0.732	0.845	Enhance Multimodal Transformer With External Label And In-Domain Pretrain: Hateful Meme Challenge Winning Solution
Pro-Cap	0.723	0.809	Pro-Cap: Leveraging a Frozen Vision-Language Model for Hateful Meme Detection
PaLI-X-VPD	-	0.892	Visual Program Distillation: Distilling Tools and Programmatic Reasoning into Vision-Language Models	-
Visual BERT COCO	0.695	0.754	The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes
Flamingo (fine-tuned)	-	0.866	Flamingo: a Visual Language Model for Few-Shot Learning
ISSUES	-	0.855	Mapping Memes to Words for Multimodal Hateful Meme Classification
HateDetectron27	0.765	0.811	Detecting Hate Speech in Memes Using Multimodal Deep Learning Approaches: Prize-winning solution to Hateful Memes Challenge	-
RGCL (CLIP)	0.788	0.870	Improving Hateful Meme Detection through Retrieval-Guided Contrastive Learning
Hate-CLIPper - Align	-	0.858	Hate-CLIPper: Multimodal Hateful Meme Classification based on Cross-modal Interaction of CLIP Features
LMM-RGCL (LLaVA-1.5-7B)	0.809	0.897	Improved Fine-Tuning of Large Multimodal Models for Hateful Meme Detection
LMM-RGCL (Qwen2-VL-2B)	0.791	0.884	Improved Fine-Tuning of Large Multimodal Models for Hateful Meme Detection

0 of 17 row(s) selected.