The CrossEntropyLoss from pytorch combines a LogSoftmax and a NLLLoss . Since you already have a Softmax layer as output activation function for ...
確定! 回上一頁