Machine Reading Comprehension on CMRC 2018 (test)
78.05EMChineseBERT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ChineseBERTModel Size=Large, Pre-training Data=Standard2021.06 | 78.05 | — | |
| RoBERTaModel Size=Large, Pre-training Data=Extended2021.06 | 77.95 | — | |
| PERTModel Size=large2022.11 | 76.8 | 90.7 | |
| LERTModel Size=large2022.11 | 75.6 | 90.9 | |
| ChineseBERTModel Size=Base, Pre-training Data=Standard2021.06 | 75.35 | — | |
| RoBERTaModel Size=Base, Pre-training Data=Extended2021.06 | 75.2 | — | |
| MacBERTModel Size=large2022.11 | 74.8 | 90.7 | |
| ERNIEModel Size=Base, Pre-training Data=Standard2021.06 | 74.7 | — | |
| RoBERTaModel Size=large2022.11 | 74.2 | 90.6 | |
| BERTModel Size=Base, Pre-training Data=Extended2021.06 | 73.95 | — | |
| ELECTRAModel Size=large2022.11 | 73.9 | 87.1 | |
| LERTModel Size=base2022.11 | 73.5 | 89.7 | |
| MacBERTModel Size=base2022.11 | 73.2 | 89.5 | |
| ELECTRAModel Size=base2022.11 | 73.1 | 87.1 | |
| PERTModel Size=base2022.11 | 72.8 | 89.2 | |
| RoBERTaModel Size=base2022.11 | 72.6 | 89.4 | |
| BERTModel Size=Base, Pre-training Data=Standard2021.06 | 71.6 | — | |
| BERTModel Size=base2022.11 | 71.4 | 87.7 | |
| BERTmodel size=base2020.04 | 70 | 87 | |
| XLNet-midmodel size=mid2020.04 | 69.3 | 89.2 | |
| XLNet-basemodel size=base2020.04 | 67 | 87.2 |