摘要
请提供需要翻译的英文内容。
代码仓库
declare-lab/RECCON
官方
pytorch
GitHub 中提及
基准测试
| 基准 | 方法 | 指标 |
|---|---|---|
| causal-emotion-entailment-on-reccon | RoBERTa Large | Macro F1: 77.06 Neg. F1: 87.89 Pos. F1: 66.23 |
| causal-emotion-entailment-on-reccon | RoBERTa Base | Macro F1: 76.51 Neg. F1: 88.74 Pos. F1: 64.28 |
| recognizing-emotion-cause-in-conversations-on | RoBERTa Base | Exact Span F1: 32.63 F1: 75.45 F1(Neg): 85.85 F1(Pos): 58.17 |
| recognizing-emotion-cause-in-conversations-on | SpanBERT | Exact Span F1: 34.64 F1: 75.71 F1(Neg): 86.02 F1(Pos): 60.00 |