Multimodal Information Extraction on M3D English
79.56Entity Recognition F1Code-MIE
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Code-MIEFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 79.56 | 72.85 | 41.49 | 50.22 | 61.03 | |
| KnowCoderFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 78.31 | 69.43 | 37.72 | 46.1 | 57.89 | |
| Code4UIEFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 77.43 | 68.35 | 35.7 | 45.14 | 56.66 | |
| CodeIEFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 77.27 | 68.95 | 35.99 | 46.26 | 57.12 | |
| CodeKGCFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 77.04 | 68.41 | 36.51 | 45.09 | 56.76 | |
| REAMOFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 74.84 | 67.06 | 34.17 | 44.85 | 55.23 | |
| Qwen3-VLFine-tuning=LoRA, Backbone=Qwen3-8B, Vision Tower=ViT-base2026.03 | 72.07 | 67.54 | 33.62 | 40.13 | 53.34 |