pytorch.sngan_projection icon indicating copy to clipboard operation
pytorch.sngan_projection copied to clipboard

Published 20 hours ago • verified publisher icon crcrpar

about why train failed

Open Johnson-yue opened this issue 6 years ago • 3 comments

Hi, train 64x64 dataset with your code . But it failed , and I test something ,like D network and loss Here is my result:

  1. sn_projection Discriminator is good, and it work fine. but wrong use in code when computed dis_fake which I think input of D should be true_label not pesudo_y . you can try it , pesudo_y is not useful at all.

  2. Hinge loss have some trouble and I was using LSGAN loss is work fine!

Johnson-yue avatar Feb 22 '19 09:02 Johnson-yue

@Johnson-yue I'm sorry for my late response and bothering you 🙇

But could you tell me whether No.2 of your results is independent of No.1?

  1. sn_projection Discriminator is good, and it work fine. but wrong use in code when computed dis_fake which I think input of D should be true_label not pesudo_y . you can try it , pesudo_y is not useful at all.
  2. Hinge loss have some trouble and I was using LSGAN loss is work fine!

crcrpar avatar Mar 01 '19 10:03 crcrpar

Sorry I reply so late, I went over the weekend.

I do not think No.2 is independent of No.1, I tested them separately!!! I am not sure my implement Hinge loss whether correctly or not, but I am sure LS loss is right!! And I am quite sure that the first problem exists.

please check it. Mybe I am wrong..... Looking forward to your reply

Johnson-yue avatar Mar 04 '19 02:03 Johnson-yue

Thank you for your answers.

I'll look into my implementation ASAP But, I've been a bit busy these days, so could you either share your implementation of Hinge and/or LS loss or send me a Pull Request that adds your patches to my loss calculation if possible?

2019年3月4日(月) 11:23 Johnson yue [email protected]:

Sorry I reply so late, I went over the weekend.

I do not think No.2 is independent of No.1, I tested them separately!!! I am not sure my implement Hinge loss whether correctly or not, but I am sure LS loss is right!! And I am quite sure that the first problem exists.

please check it. Mybe I am wrong..... Looking forward to your reply

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/crcrpar/pytorch.sngan_projection/issues/23#issuecomment-469096286, or mute the thread https://github.com/notifications/unsubscribe-auth/APcP0xRZOH9LrWodMywgo6yxEjWj0Q2fks5vTIOjgaJpZM4bJJr_ .

crcrpar avatar Mar 04 '19 02:03 crcrpar