Relation Extraction On Funsd
評価指標
F1
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | F1 | Paper Title | Repository |
|---|---|---|---|
| LayoutLMv3 large EM + BBO + RSF | 90.81 | A LayoutLMv3-Based Model for Enhanced Relation Extraction in Visually-Rich Documents | - |
| LayoutLMv3 large | 80.35 | LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking | |
| LayoutLMv3 large | 80.35 | GeoLayoutLM: Geometric Pre-training for Visual Information Extraction | |
| LayoutLMv2 large | 70.57 | LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding | |
| BROS | 77.01 | BROS: A Pre-trained Language Model Focusing on Text and Layout for Better Key Information Extraction from Documents | |
| GeoLayoutLM | 89.45 | GeoLayoutLM: Geometric Pre-training for Visual Information Extraction | |
| RORE (GeoLayoutLM) | 88.46 | Modeling Layout Reading Order as Ordering Relations for Visually-rich Document Understanding | |
| LayoutLM | 42.83 | LayoutLM: Pre-training of Text and Layout for Document Image Understanding | |
| TPP (LayoutMask) | 79.20 | Reading Order Matters: Information Extraction from Visually-rich Documents by Token Path Prediction |
0 of 9 row(s) selected.