I see image-classification models from torchvision package don't have a softmax layer as final layer. For instance, the following snippet ...
確定! 回上一頁