对于我的 MT 系统,我使用https://huggingface.co/metrics/rouge计算了 ROUGE-L 值。输出如下所示。大多数论文都报告了一个 ROUGE-L 值,我也想做同样的事情。但是,我的输出如下所示,我不确定要报告哪个值?我应该报告低、中还是高?和进动或召回或F-measure?
'rougeL': AggregateScore(low=Score(precision=0.34535176087958586, recall=0.36969750745470553, fmeasure=0.33939664257593155), mid=Score(precision=0.40405631462907, recall=0.41156890941875457, fmeasure=0.3835437703820411), high=Score(precision=0.4648738881460244, recall= 0.4597817743860313, fmeasure=0.43226391587929297))