Hi all,
when I used DNN-Operations of math kernel library on win7, visual studio2015, it worked well.
convolution and maxpooling functions get same results, but faster.
but, the convolution functions get different value changed on win10,
do you have any ideas?
Thanks.