Semantic Parsing on Prague Tectogrammatical Graphs
93.58Properties F1RobeCzech
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RobeCzech2021.05 | 93.58 | 96.57 | 97.97 | 84.92 | 78.29 | 92.36 | |
| ÚFAL+XLM-R large2021.05 | 93.56 | 96.23 | 97.86 | 84.61 | 78.62 | 92.24 | |
| XLM-R large2021.05 | 93.31 | 96.42 | 97.92 | 84.46 | 77.89 | 92.11 | |
| XLM-R base2021.05 | 93.12 | 96.09 | 97.6 | 83.03 | 76.16 | 91.55 | |
| Slavic BERT2021.05 | 92.91 | 95.92 | 97.51 | 82.48 | 75.08 | 91.27 | |
| RobeCzechembeddings=frozen2021.05 | 92.76 | 95.85 | 97.41 | 82.6 | 74.95 | 91.23 | |
| Czert2021.05 | 92.69 | 95.72 | 97.23 | 80.91 | 72.37 | 90.66 | |
| mBERT2021.05 | 92.6 | 95.72 | 97.2 | 80.77 | 72.83 | 90.62 | |
| Hitachi+mBERT2021.05 | 91.48 | 87.69 | 93.99 | 76.9 | 66.07 | 87.35 | |
| HIT-SCIR+mBERT2021.05 | 79.01 | 84.14 | 92.34 | 64.96 | 47.68 | 77.93 | |
| HUJI-KU+mBERT2021.05 | 72.44 | — | 72.1 | 44.91 | — | 58.49 |