Question Matching on LCQMC (dev)
90.9AccuracyERNIE 2.0
Evaluation results
| Method | Links | |
|---|---|---|
| ERNIE 2.0Model Version=ERNIE 2.0, Model Scale=BASE2019.07 | 90.9 | |
| ERNIE 2.0Model Version=ERNIE 2.0, Model Scale=LARGE2019.07 | 90.9 | |
| ELECTRA-largemodel size=large2020.04 | 90.7 | |
| MacBERT-largemodel size=large2020.04 | 90.6 | |
| MacBERTModel Scale=Large, Pre-trained on extended data=false2021.06 | 90.6 | |
| ChineseBERTModel Scale=Large, Pre-trained on extended data=false2021.06 | 90.5 | |
| ROBERTa-wwm-ext-largemodel size=large2020.04 | 90.4 | |
| RoBERTaModel Scale=Large, Pre-trained on extended data=true2021.06 | 90.4 | |
| ELECTRA-basemodel size=base2020.04 | 90.2 | |
| ERNIEModel Scale=Base, Pre-trained on extended data=false2021.06 | 89.8 | |
| ChineseBERTModel Scale=Base, Pre-trained on extended data=false2021.06 | 89.8 | |
| ERNIE 1.0Model Version=ERNIE 1.0, Model Scale=BASE2019.07 | 89.7 | |
| BERT-wwm-extmodel size=base2020.04 | 89.6 | |
| BERTModel Scale=Base, Pre-trained on extended data=true2021.06 | 89.6 | |
| MacBERT-basemodel size=base2020.04 | 89.5 | |
| MacBERTModel Scale=Base, Pre-trained on extended data=false2021.06 | 89.5 | |
| BERTmodel size=base2020.04 | 89.4 | |
| BERT-wwmmodel size=base2020.04 | 89.4 | |
| BERTModel Scale=Base, Pre-trained on extended data=false2021.06 | 89.4 | |
| ROBERTa-wwm-extmodel size=base2020.04 | 89 | |
| RoBERTaModel Scale=Base, Pre-trained on extended data=true2021.06 | 89 | |
| BERTModel Version=BERT, Model Scale=BASE2019.07 | 88.8 | |
| ERNIE2019.04 | 0.897 | |
| BERT2019.04 | 0.888 |