NLLLoss ) with log-softmax (tensor. I;m using Mask RCNN and trying to change SGD, which is an existing optimizer, to Adam but the loss value is output as nan ...
確定！ 回上一頁
