Text Classification on Arxiv BERT-base (test)
90.5AccuracyAdapterFusion
Evaluation results
| Method | Links | |
|---|---|---|
| AdapterFusionNum Params=182 M, Backbone=BERT-base2022.03 | 90.5 | |
| ADA (K=4) - TransRateNum Params=128 M, Backbone=BERT-base2022.03 | 89.8 | |
| AdaptersNum Params=182 M, Backbone=BERT-base2022.03 | 87.2 | |
| ADA (K=4) - LEEPNum Params=128 M, Backbone=BERT-base2022.03 | 86.7 | |
| ADA (K=1)Num Params=120.8 M, Backbone=BERT-base2022.03 | 77.2 |