Dependency Parsing on TB2 (test)
83.82UASspaCy-XLM-RoBERTa
Evaluation results
| Method | Links | ||
|---|---|---|---|
| spaCy-XLM-RoBERTaTraining Data=TB2, Backbone=XLM-RoBERTa2022.01 | 83.82 | 79.39 | |
| Liu et al. (2018)Training Data=TB2, Model Configuration=Ensemble2022.01 | 83.4 | 79.4 | |
| Liu et al. (2018)Training Data=TB2, Model Configuration=Distillation2022.01 | 82.1 | 77.9 | |
| StanzaTraining Data=TB2 + EWT2022.01 | 82.1 | 77.6 | |
| Dozat et al. (2017)Training Data=TB22022.01 | 81.8 | 77.7 | |
| Kong et al. (2014)Training Data=TB22022.01 | 81.4 | 76.9 | |
| spaCy-XLM-RoBERTaTraining Data=TB2 + EWT, Backbone=XLM-RoBERTa2022.01 | 81.02 | 75.43 | |
| Ballesteros et al. (2015)Training Data=TB22022.01 | 80.2 | 75.7 | |
| StanzaTraining Data=TB22022.01 | 79.28 | 74.34 | |
| spaCy-BERTweetTraining Data=TB2, Backbone=BERTweet2022.01 | 76.32 | 71.72 | |
| spaCy-BERTweetTraining Data=TB2 + EWT, Backbone=BERTweet2022.01 | 76.18 | 69.28 | |
| spaCyTraining Data=TB2 + EWT2022.01 | 72.06 | 63.84 | |
| spaCyTraining Data=TB22022.01 | 66.93 | 58.79 |