Chunking On Penn Treebank
Métriques
F1 score
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | F1 score |
|---|---|
| contextual-string-embeddings-for-sequence | 96.72 |
| ncrf-an-open-source-neural-sequence-labeling | 95.06 |
| Modèle 3 | 95.15 |
| a-joint-many-task-model-growing-a-neural | 95.77 |
| deep-multi-task-learning-with-low-level-tasks | 95.57 |
| bidirectional-lstm-crf-models-for-sequence | 94.46 |
| learning-better-internal-structure-of-words | 95.29 |
| automated-concatenation-of-embeddings-for-1 | 97.3 |