When using packed padded sequences, we need to tell PyTorch how long the ... We use the pad token index to create the masks, by creating a mask tensor that ...
確定! 回上一頁