WGAN-gp loss keeps going large

[haonanhe]

Hello, I've implemented your code on my own dataset. However, the d_loss decreases from 10(which equals to lambda) to a very small negative number(like -10000), the wasserstein distance keeps going to order of million, and the gradient penalty changes from 10 to 0 and then goes to order of thousand. I've worked on this problem for several days but I still can't solve it. Can anyone help me with this?
@caogang

[yallien]

Hello,I've met the same problem as yours,did u find out where the problem is?

[yifanjiang19]

@yallien @supermarian Are you using pytorch with version >= 1.4?

[mengxiangxiang414]

@haonanhe @yallien hello,did u solve this problem?

[hangexuexi]

how about the wasserstein distance at the start of the train, it is very smaller than lambda?