Finally, for slightly better performance, we perform the softmax while computing the loss ... 2 http://deeplearning.net/software/theano/; http://torch.ch/; ...
確定! 回上一頁