Important: The padding item should be mask when calculating CRF Loss, ... Consider a Conv2D layer: it can only be called on a single input tensor of rank 4.
確定! 回上一頁