You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
我在论文中看到:The network defines a trasnformation which transforms an raw input text x to r-dimensional deep representation h.
这是不是就意味着最后的h是一个r维的向量呢?
因为注意到文章在超参数一节中写道CNN的两层中共使用了两次Folding,我疑惑的是,48行的矩阵输入经过两次folding是不是就应该变成12行矩阵呢,想请问最后是怎么成为h向量的呢?