NLP Classification on GLUE
87.7Average Test Accuracy (S1)Individual
Evaluation results
| Method | Links | ||
|---|---|---|---|
| IndividualBase Model=BERT2022.10 | 87.7 | 0 | |
| EXSSNETBase Model=BERT, Classification Head=CNN-based2022.10 | 80.5 | 0 | |
| MultitaskBase Model=BERT2022.10 | 79.9 | 0 | |
| MultiAdaptBERTBase Model=BERT2022.10 | 78.5 | 0 | |
| SSNETBase Model=BERT, Classification Head=CNN-based2022.10 | 78.4 | 3.6 | |
| SupSupBase Model=BERT, Classification Head=CNN-based2022.10 | 78.3 | 0 | |
| ReplayBase Model=BERT2022.10 | 77.7 | 4.8 | |
| AdaptBERT + ReplayBase Model=BERT2022.10 | 76.8 | 3.8 | |
| Prompt TuningBase Model=BERT2022.10 | 76.3 | 0 | |
| IDBRBase Model=BERT2022.10 | 73 | 6.8 | |
| RegularizationBase Model=BERT2022.10 | 72.5 | 8.8 | |
| RandomBase Model=BERT2022.10 | 33.3 | — | |
| AdaptBERT + FTBase Model=BERT2022.10 | 24.7 | 53.4 | |
| FTBase Model=BERT2022.10 | 14.1 | 86 |