Is there a better way to do this? How to pad tensor with zeros, without creating new tensor object? I need inputs to be of the same batchsize ...
確定! 回上一頁