Harmful meme detection on Harm-P (test)
90.73AccuracyEXPLAINHM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EXPLAINHM2024.01 | 90.73 | 90.72 | |
| MOMENTA2024.01 | 89.84 | 88.26 | |
| Pro-Cap2024.01 | 89.32 | 87.91 | |
| MaskPrompt2024.01 | 88.17 | 87.09 | |
| VILBERT CCPre-training dataset=CC2024.01 | 87.25 | 86.03 | |
| VisualBERT COCOPre-training dataset=COCO2024.01 | 86.8 | 86.07 | |
| MMBT2024.01 | 82.54 | 80.23 | |
| Text BERT2024.01 | 80.12 | 78.35 | |
| Late Fusion2024.01 | 78.26 | 78.5 | |
| Image-Region2024.01 | 73.14 | 72.77 |