sohelmsc

So basically you have to use .item() as follows instead of using .data[0]. This is actually for new version of torch. D_cost = fake_loss_48.item() + fake_loss_32.item() + fake_loss_16.item() + fake_loss_0.item()

I trained the model for different datasets. Can you give me the scenario where you are getting the error? Thanks On Tue, Jul 23, 2019 at 10:17 PM zhangbaijin wrote:...