Numerical Reasoning Question Answering on FinQA v1 (dev)
72.91Execution AccuracyAPOLLO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| APOLLOProtocol=Fine-tuning, Ensemble=true2022.12 | 72.91 | 70.83 | |
| APOLLOProtocol=Fine-tuning2022.12 | 69.7 | 65.91 | |
| DyRRenProtocol=Fine-tuning, Backbone=ROBERTa-large2022.12 | 66.82 | 63.87 | |
| ELASTICProtocol=Fine-tuning, Backbone=ROBERTa-large2022.12 | 65 | 61 | |
| FinQANetProtocol=Fine-tuning, Backbone=ROBERTa-large2022.12 | 61.22 | 58.05 | |
| Retriever+NeRdProtocol=Fine-tuning, Backbone=ROBERTa-large2022.12 | 47.53 | 45.37 | |
| LongformerProtocol=Fine-tuning, Backbone=ROBERTa-large2022.12 | 23.83 | 22.56 |