Question Answering On Squad20 Dev
Métriques
EM
F1
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | EM | F1 |
|---|---|---|
| albert-a-lite-bert-for-self-supervised | 76.1 | 79.1 |
| roberta-a-robustly-optimized-bert-pretraining | 86.5 | 89.4 |
| albert-a-lite-bert-for-self-supervised | 79.0 | 82.1 |
| xlnet-generalized-autoregressive-pretraining | 87.9 | 90.6 |
| read-verify-machine-reading-comprehension | 72.3 | 74.8 |
| semantics-aware-bert-for-language | 80.9 | 83.6 |
| spanbert-improving-pre-training-by | - | 86.8 |
| sg-net-syntax-guided-machine-reading | 85.1 | 87.9 |
| 190910351 | 69.9 | 73.4 |
| dice-loss-for-data-imbalanced-nlp-tasks | 87.65 | 89.51 |
| albert-a-lite-bert-for-self-supervised | 83.1 | 85.9 |
| u-net-machine-reading-comprehension-with | 70.3 | 74.0 |
| albert-a-lite-bert-for-self-supervised | 85.1 | 88.1 |