Multi-label classification on DPM PCL Subtask 2 (test)
62.78Unbalanced Power Relations F1Ensemble 1.0
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Ensemble 1.0Official Ranking=true, Fine-tuning=true2022.03 | 62.78 | 54.76 | 34.15 | 34.19 | 33.33 | 50.82 | 8.7 | 39.82 | |
| Ensemble 2.0Fine-tuning=true2022.03 | 61.4 | 55.81 | 39.13 | 34.15 | 43.84 | 49.61 | 35.71 | 45.66 | |
| BERT-PCLFine-tuning=true, Pre-trained backbone=BERT2022.03 | 58.9 | 50.55 | 42.86 | 28.07 | 40 | 49.24 | 33.33 | 43.28 | |
| RoBERTaFine-tuning=true, Pre-trained backbone=RoBERTa2022.03 | 57.41 | 50.6 | 26.83 | 26.79 | 40.51 | 48.28 | 14.81 | 37.89 | |
| ERNIE 2.0Fine-tuning=true, Pre-trained backbone=ERNIE 2.02022.03 | 56.5 | 55.88 | 25.97 | 16.49 | 28.07 | 45.9 | 0 | 32.69 | |
| BERTFine-tuning=true, Pre-trained backbone=BERT2022.03 | 51.52 | 40.62 | 20.29 | 6.98 | 8.7 | 42.48 | 0 | 24.37 | |
| ALBERTFine-tuning=true, Pre-trained backbone=ALBERT2022.03 | 51.31 | 41.79 | 16.39 | 9.64 | 9.52 | 40 | 0 | 24.09 | |
| RandomFine-tuning=false2022.03 | 10.82 | 2.23 | 3.03 | 4.21 | 1.91 | 5.98 | 1.22 | 4.2 |