thinkInJava33

How long does it take to train on imagenet？It takes much more time than it mentioned in the paper

> Hi @tonystark940501, > I train ImageNet dataset with 2 K80 GPU, not slow, Each epoch use ~3 hours, > > 2019-01-22 13:37:49,194 epoch 0 lr 1.000000e-01 > 2019-01-22 13:38:07,415...

我也是 ，有反馈的途径吗

do you get the same result on CIFAR100 as mentioned in the paper

@cvtower i follow your suggustions,but i still can not reproduce the result on cifar100.

@cvtower thank you,i am wating for it,and i hope it won't be too long.

224 maybe

losses.update(loss.data[0], input.size(0))->losses.update(loss.item(),input.size(0))

@jyh2986 really need your help

i took me 3 hours per epoch with 2 1080ti