Metric: F1 (higher is better)
| # | Model↕ | F1▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | CamemBERT-Large | 92.2 | No | FQuAD: French Question Answering Dataset | 2020-02-14 | - |
| 2 | XLM-RoBERTa-Large | 89.5 | No | FQuAD: French Question Answering Dataset | 2020-02-14 | - |
| 3 | CamemBERT-Base | 88.4 | No | FQuAD: French Question Answering Dataset | 2020-02-14 | - |
| 4 | Camembert-Base-SquadFR-Fquad-Piaf | 88.4 | Yes | Project PIAF: Building a Native French Question-... | 2020-07-02 | Code |
| 5 | XLM-RoBERTa-Base | 85.9 | No | FQuAD: French Question Answering Dataset | 2020-02-14 | - |
| 6 | Fmikaelian-Camembert-Base-Fquad | 85.6 | No | - | - | - |
| 7 | LePetit | 70.71 | No | On the importance of pre-training data volume fo... | 2020-10-08 | - |