caffe编译prototxt文件时出现错误(SIGMOID CROSS) 5
I032115:34:28.46763318044layer_factory.hpp:77]CreatinglayercropI032115:34:28.46763318...
I0321 15:34:28.467633 18044 layer_factory.hpp:77] Creating layer cropI0321 15:34:28.467633 18044 net.cpp:100] Creating Layer cropI0321 15:34:28.467633 18044 net.cpp:444] crop <- score-dsn6-upI0321 15:34:28.467633 18044 net.cpp:444] crop <- data_data_0_split_1I0321 15:34:28.467633 18044 net.cpp:418] crop -> upscore-dsn6I0321 15:34:28.467633 18044 net.cpp:150] Setting up cropI0321 15:34:28.467633 18044 net.cpp:157] Top shape: 1 1 300 400 (120000)I0321 15:34:28.467633 18044 net.cpp:165] Memory required for data: 424776080I0321 15:34:28.467633 18044 layer_factory.hpp:77] Creating layer upscore-dsn6_crop_0_splitI0321 15:34:28.467633 18044 net.cpp:100] Creating Layer upscore-dsn6_crop_0_splitI0321 15:34:28.483258 18044 net.cpp:444] upscore-dsn6_crop_0_split <- upscore-dsn6I0321 15:34:28.483258 18044 net.cpp:418] upscore-dsn6_crop_0_split -> upscore-dsn6_crop_0_split_0I0321 15:34:28.483258 18044 net.cpp:418] upscore-dsn6_crop_0_split -> upscore-dsn6_crop_0_split_1I0321 15:34:28.483258 18044 net.cpp:150] Setting up upscore-dsn6_crop_0_splitI0321 15:34:28.483258 18044 net.cpp:157] Top shape: 1 1 300 400 (120000)I0321 15:34:28.483258 18044 net.cpp:157] Top shape: 1 1 300 400 (120000)I0321 15:34:28.483258 18044 net.cpp:165] Memory required for data: 425736080I0321 15:34:28.483258 18044 layer_factory.hpp:77] Creating layer loss-dsn6I0321 15:34:28.483258 18044 net.cpp:100] Creating Layer loss-dsn6I0321 15:34:28.483258 18044 net.cpp:444] loss-dsn6 <- upscore-dsn6_crop_0_split_0I0321 15:34:28.483258 18044 net.cpp:444] loss-dsn6 <- label_data_1_split_0I0321 15:34:28.483258 18044 net.cpp:418] loss-dsn6 -> loss-dsn6F0321 15:34:28.483258 18044 sigmoid_cross_entropy_loss_layer.cpp:23] Check failed: bottom[0]->count() == bottom[1]->count() (120000 vs. 1) SIGMOID_CROSS_ENTROPY_LOSS layer inputs must have the same count.*** Check failure stack trace: ***这个不算太懂sigmoid这个函数还有这个错误究竟该如何去掉?
展开
1个回答
推荐律师服务:
若未解决您的问题,请您详细描述您的问题,通过百度律临进行免费专业咨询