revise：After the operation of two convolution layers and interaction layers a richer context feature FM can be obtained by aggregating the features of the two subtasks
By applying two convolution layers and interaction layers, we can aggregate the features of the two subtasks and obtain a more comprehensive context feature FM.
原文地址: https://www.cveoy.top/t/topic/iV6f 著作权归作者所有。请勿转载和采集!