I agree with you that the documentation for nn.NLLLoss() is far from ideal, but I think we can clarify your problem here, firstly, by clarifying that ...
確定! 回上一頁