thuml/DAN

Issues

• There occurring an error when I run cmake cmd,and the error details are available in the CMakeOutput.log and CMakeError.log following in the appendix.

  #13 opened 4 years ago by GaoZiqiang
  0

• The value of mmd loss is always negative . could anyone explain this issue?

  #6 opened 7 years ago by yumaofan
  2

• Results are not same as per paper.

  #9 opened 7 years ago by redhat12345
  1

• Check failed: false Unknown filler name: identity

  #10 opened 7 years ago by mahfuj9346449
  1

• What's happened after optimization done.

  #11 opened 7 years ago by wujun
  2

• 您好，请教一个问题，在《Learning Transferable Features with Deep Adaptation Networks》这篇论文中，在求解网络参数$theta$时，论文将第L层参数的梯度表示为：J（Zi）对第L层参数的偏导+gk（zli）对第L层参数的偏导，如果按照本文CNN的目标函数，第二项不应该是从L6-L8三层的gk和么？这里单独拿出来第L层的gk应该怎么理解？希望能得到您的指点，非常感谢！

  #12 opened 7 years ago by zhangxiaozao
  0

• How to port your modifications to latest caffe

  #7 opened 7 years ago by wujun
  0

• JAN-A on Resnet50 : is loss_weight -0.3 for "fc7_mmd_loss"?

  #2 opened 7 years ago by YifeiAI
  0

• Cannot find the Office dataset

  #8 opened 7 years ago by jindongwang
  1

• JMMD loss value is zero!

  #5 opened 7 years ago by swamiviv
  1

• line 617 should be "target_softmax", rather than "target_label";

  #4 opened 7 years ago by guicunbin
  1

• Negative MMD loss

  #3 opened 7 years ago by insmod-he
  2