Text Classification on BQ Corpus (dev)
86.7Accuracy (Max)ELECTRA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ELECTRAScale=large2022.03 | 86.7 | 86.2 | |
| ELECTRAModel size=large2022.11 | 86.7 | 86.2 | |
| LERTModel size=large2022.11 | 86.6 | 86 | |
| BERTScale=base2022.03 | 86.4 | 85.5 | |
| BERTModel size=base2022.11 | 86.4 | 85.5 | |
| RoBERTaScale=large2022.03 | 86.3 | 85.7 | |
| PERTScale=large2022.03 | 86.3 | 85.8 | |
| RoBERTaModel size=large2022.11 | 86.3 | 85.7 | |
| PERTModel size=large2022.11 | 86.3 | 85.8 | |
| MacBERTScale=large2022.03 | 86.2 | 85.7 | |
| MacBERTModel size=large2022.11 | 86.2 | 85.7 | |
| RoBERTaScale=base2022.03 | 86 | 85.4 | |
| MacBERTScale=base2022.03 | 86 | 85.5 | |
| MacBERTModel size=base2022.11 | 86 | 85.2 | |
| LERTModel size=base2022.11 | 85.9 | 85.6 | |
| RoBERTaModel size=base2022.11 | 85.3 | 84.8 | |
| PERTScale=base2022.03 | 84.9 | 84.8 | |
| PERTModel size=base2022.11 | 84.9 | 84.8 | |
| ELECTRAScale=base2022.03 | 84.8 | 84.7 | |
| ELECTRAModel size=base2022.11 | 84.8 | 84.7 |