Relation Classification on TACRED K=16 original (low-resource)
36.6Micro F1GenPT (RoBERTa)
Evaluation results
| Method | Links | |
|---|---|---|
| GenPT (RoBERTa)Learning Paradigm=Prompt Tuning, Backbone=RoBERTa2022.10 | 36.6 | |
| NLILearning Paradigm=Fine-tuning2022.10 | 35.6 | |
| GenPT (BART)Learning Paradigm=Prompt Tuning, Backbone=BART2022.10 | 35.6 | |
| KnowPromptLearning Paradigm=Prompt Tuning, Backbone=RoBERTa2022.10 | 35.4 | |
| GenPT (T5)Learning Paradigm=Prompt Tuning, Backbone=T52022.10 | 33.3 | |
| TYP MarkerLearning Paradigm=Fine-tuning, Backbone=RoBERTa2022.10 | 32 | |
| PTRLearning Paradigm=Prompt Tuning, Backbone=RoBERTa2022.10 | 30.7 | |
| TANLLearning Paradigm=Fine-tuning, Backbone=T52022.10 | 27.6 | |
| GDPNetLearning Paradigm=Fine-tuning2022.10 | 22.5 | |
| LUKELearning Paradigm=Fine-tuning, Backbone=LUKE2022.10 | 21.5 | |
| SpanBERTLearning Paradigm=Fine-tuning, Backbone=SpanBERT2022.10 | 17.5 |