search

Jianlong-Fu / Recurrent-Attention-CNN

225 stars 45 forks source link

The question about training with rank loss . #16

Open suhuijia opened 6 years ago

suhuijia commented 6 years ago

@Jianlong-Fu Hi, Jianlong, I have your train_val_cnn.prototxt and code ( include rank_loss2_layer), but when I make triaining, it do not work.. Has been staying such as: I0524 11:50:19.434263 13879 net.cpp:242] This network produces output accuracy I0524 11:50:19.434267 13879 net.cpp:242] This network produces output loss1 I0524 11:50:19.434268 13879 net.cpp:242] This network produces output loss2 I0524 11:50:19.434273 13879 net.cpp:242] This network produces output loss3 I0524 11:50:19.434274 13879 net.cpp:242] This network produces output rank I0524 11:50:19.434320 13879 net.cpp:255] Network initialization done. I0524 11:50:19.434532 13879 solver.cpp:56] Solver scaffolding done. I0524 11:50:19.438776 13879 caffe.cpp:249] Starting Optimization I0524 11:50:19.438781 13879 solver.cpp:273] Solving RA_CNN I0524 11:50:19.438783 13879 solver.cpp:274] Learning Rate Policy: step

What's the problem, please? I really need you help. Thank you very much!

jackshaw commented 6 years ago

Hi,@suhuijia,could please send me the train_val_cnn.prototxt?My email address is 2689346162@qq.com.

ghost commented 5 years ago

Hi @suhuijia , could you also send me the train_val_cnn.prototxt? My email is byz5@duke.edu.