Dependency Parsing On Universal Dependencies
评估指标
LAS
UAS
评测结果
各个模型在此基准测试上的表现结果
| Paper Title | Repository | |||
|---|---|---|---|---|
| UDPipe 2.0 + mBERT + FLAIR | 84.60 | 87.64 | Evaluating Contextualized Embeddings on 54 Languages in POS Tagging, Lemmatization and Dependency Parsing | - |
| UDify | 80.43 | 85.69 | 75 Languages, 1 Model: Parsing Universal Dependencies Universally | |
| HIT-SCIR | 75.84 | - | Towards Better UD Parsing: Deep Contextualized Word Embeddings, Ensemble, and Treebank Concatenation | |
| Stanford+ | 74.16 | 62.08 | Universal Dependency Parsing from Scratch | |
| TurkuNLP | 73.28 | 60.99 | Turku Neural Parser Pipeline: An End-to-End System for the CoNLL 2018 Shared Task | - |
| UDPipe 2.0 | 73.11 | 61.25 | UDPipe 2.0 Prototype at CoNLL 2018 UD Shared Task | - |
0 of 6 row(s) selected.