TimeDial

该数据集主要关注对话式、非知识性背景下的情境。在其他信息中，我们注意到TSM模型在TimeDIAL-0上的表现最为出色，这很可能是由于DailyDialog训练数据集的规模较大。该任务的目的是进行时间跨度评分。This dataset primarily focuses on conversational, non-informational contextual scenarios. Notably, we observed that the TSM model achieved the best performance on TimeDIAL-0, which is likely attributed to the larger scale of the DailyDialog training dataset. The objective of this task is temporal span scoring.