Multi-label Emotion Classification on Arabic dataset (test)
54.88F1-MacroSLCL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SLCLloss_type=Stepwise Label Contrastive Loss2022.12 | 54.88 | 66.37 | 54.65 | |
| SCLloss_type=Strict Contrastive Loss2022.12 | 54.27 | 66.71 | 55.43 | |
| ICLloss_type=Intra-label Contrastive Loss2022.12 | 54.26 | 66.13 | 54.17 | |
| JSCLloss_type=Jaccard Similarity Contrastive Loss2022.12 | 54.08 | 66 | 54.14 | |
| JSPCLloss_type=Jaccard Similarity Probability Contrastive Loss2022.12 | 53.7 | 65.86 | 53.98 | |
| SpanEmo2022.12 | 53.63 | 65.81 | 53.94 | |
| HEF2022.12 | 50.2 | 63.1 | 51.2 | |
| BERTbaseBackbone=BERT-base2022.12 | 47.7 | 65 | 52.3 | |
| EMA2022.12 | 46.1 | 61.8 | 48.9 | |
| Tw-StAR2022.12 | 44.6 | 59.7 | 46.5 |