Relation Classification on Re-TACRED K=8 low-resource re-evaluated
57.1Micro F1GenPT (RoBERTa)
Evaluation results
| Method | Links | |
|---|---|---|
| GenPT (RoBERTa)Learning Paradigm=Prompt Tuning, Backbone=RoBERTa2022.10 | 57.1 | |
| KnowPromptLearning Paradigm=Prompt Tuning, Backbone=RoBERTa2022.10 | 55.3 | |
| GenPT (T5)Learning Paradigm=Prompt Tuning, Backbone=T52022.10 | 54.7 | |
| GenPT (BART)Learning Paradigm=Prompt Tuning, Backbone=BART2022.10 | 52.9 | |
| PTRLearning Paradigm=Prompt Tuning, Backbone=RoBERTa2022.10 | 51.5 | |
| TYP MarkerLearning Paradigm=Fine-tuning, Backbone=RoBERTa2022.10 | 44.8 | |
| TANLLearning Paradigm=Fine-tuning, Backbone=T52022.10 | 26.7 | |
| GDPNetLearning Paradigm=Fine-tuning2022.10 | 18.8 | |
| SpanBERTLearning Paradigm=Fine-tuning, Backbone=SpanBERT2022.10 | 14.2 | |
| LUKELearning Paradigm=Fine-tuning, Backbone=LUKE2022.10 | 14.1 |