CrossEntropyLoss ) with logits output in the forward() method, or you can use negative log-likelihood loss (tensor.nn.
確定! 回上一頁