Natural Language Inference on XNLI (dev)
82.7AccuracyChineseBERT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ChineseBERTModel Size=Large2021.06 | 82.7 | — | — | |
| ERNIE 2.0Model Version=ERNIE 2.0, Model Scale=LARGE2019.07 | 82.6 | — | — | |
| MacBERTModel Size=Large2021.06 | 82.4 | — | — | |
| MacBERT-largeModel Size=large2019.06 | 82.4 | — | — | |
| RoBERTaModel Size=Large, Pre-training Data=Extended2021.06 | 82.1 | — | — | |
| RoBERTa-wwm-ext-largeModel Size=large, Whole Word Masking=true, Extended Training=true2019.06 | 82.1 | — | — | |
| ELECTRA-largeModel Size=large2019.06 | 81.5 | — | — | |
| ERNIE 2.0Model Version=ERNIE 2.0, Model Scale=BASE2019.07 | 81.2 | — | — | |
| ChineseBERTModel Size=Base2021.06 | 80.5 | — | — | |
| MacBERTModel Size=Base2021.06 | 80.3 | — | — | |
| MacBERT-baseModel Size=base2019.06 | 80.3 | — | — | |
| RoBERTaModel Size=Base, Pre-training Data=Extended2021.06 | 80 | — | — | |
| RoBERTa-wwm-extModel Size=base, Whole Word Masking=true, Extended Training=true2019.06 | 80 | — | — | |
| ERNIE2019.04 | 79.9 | — | — | |
| ERNIE 1.0Model Version=ERNIE 1.0, Model Scale=BASE2019.07 | 79.9 | — | — | |
| ERNIEModel Size=Base2021.06 | 79.7 | — | — | |
| BERTModel Size=Base, Pre-training Data=Extended2021.06 | 79.4 | — | — | |
| BERT-wwm-extModel Size=base, Whole Word Masking=true, Extended Training=true2019.06 | 79.4 | — | — | |
| BERTModel Size=Base2021.06 | 79 | — | — | |
| BERT-wwmModel Size=base, Whole Word Masking=true2019.06 | 79 | — | — | |
| BERT2019.04 | 78.1 | — | — | |
| BERTModel Version=BERT, Model Scale=BASE2019.07 | 78.1 | — | — | |
| ELECTRA-baseModel Size=base2019.06 | 77.9 | — | — | |
| BERTModel Size=base2019.06 | 77.8 | — | — | |
| BERTScale=base2022.03 | — | 79.4 | 78.6 | |
| BERTModel size=base2022.11 | — | 79.4 | 78.6 | |
| ELECTRAScale=base2022.03 | — | 77.9 | 77 | |
| ELECTRAScale=large2022.03 | — | 81.5 | 80.8 | |
| ELECTRAModel size=base2022.11 | — | 77.9 | 77 | |
| ELECTRAModel size=large2022.11 | — | 81.5 | 80.8 | |
| LERTModel size=base2022.11 | — | 80.2 | 79.5 | |
| LERTModel size=large2022.11 | — | 81.7 | 81.2 | |
| MacBERTScale=base2022.03 | — | 80.3 | 79.7 | |
| MacBERTScale=large2022.03 | — | 82.4 | 81.8 | |
| MacBERTModel size=base2022.11 | — | 80.3 | 79.7 | |
| MacBERTModel size=large2022.11 | — | 82.4 | 81.8 | |
| PERTScale=base2022.03 | — | 78.8 | 78.1 | |
| PERTScale=large2022.03 | — | 81 | 80.4 | |
| PERTModel size=base2022.11 | — | 78.8 | 78.1 | |
| PERTModel size=large2022.11 | — | 81 | 80.4 | |
| RoBERTaScale=base2022.03 | — | 80 | 79.2 | |
| RoBERTaScale=large2022.03 | — | 82.1 | 81.3 | |
| RoBERTaModel size=base2022.11 | — | 80 | 79.2 | |
| RoBERTaModel size=large2022.11 | — | 82.1 | 81.3 |