Task Oriented Dialogue Systems On Kvret
Métriques
BLEU
Entity F1
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | BLEU | Entity F1 |
|---|---|---|
| amalgamating-knowledge-from-two-teachers-for | 17.35 | 55.38 |
| mem2seq-effectively-incorporating-knowledge | 12.6 | 33.4 |
| dynamic-memory-enhanced-transformer-for-end | 17.3 | 63.6 |
| global-to-local-memory-pointer-networks-for | 14.79 | 59.97 |
| sequence-to-sequence-learning-for-task | 12.7 | 51.9 |
| unifiedskg-unifying-and-multi-tasking | - | 70.07 |
| entity-consistent-end-to-end-task-oriented | 13.9 | 53.7 |
| key-value-retrieval-networks-for-task | 13.2 | 48.0 |
| dynamic-fusion-network-for-multi-domain-end | 15.2 | 62.5 |
| a-template-guided-hybrid-pointer-network-for-1 | 12.8 | 37.8 |