Chunking On Penn Treebank
평가 지표
F1 score
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | F1 score | Paper Title | Repository |
|---|---|---|---|
| Flair embeddings | 96.72 | Contextual String Embeddings for Sequence Labeling | |
| NCRF++ | 95.06 | NCRF++: An Open-source Neural Sequence Labeling Toolkit | - |
| Suzuki and Isozaki | 95.15 | - | - |
| JMT | 95.77 | A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks | - |
| Low supervision | 95.57 | - | - |
| BI-LSTM-CRF (Senna) (ours) | 94.46 | Bidirectional LSTM-CRF Models for Sequence Tagging | - |
| IntNet + BiLSTM-CRF | 95.29 | Learning Better Internal Structure of Words for Sequence Labeling | - |
| ACE | 97.3 | Automated Concatenation of Embeddings for Structured Prediction | - |
0 of 8 row(s) selected.