my loss can not converge

[zjz5250]

@wy1iu hi， 1 have two problems： 1）when i train the model with “type: SINGLE”，the loss can not converge，it is always 9.xx 2）then i use the model trained by 1），to train a new model，but the loss become bigger and bigger i don't know why，hope your help。 my sovlver： base_lr: 0.1 momentum: 0.9 lr_policy: "multistep" stepvalue: 16000 stepvalue: 24000 stepvalue: 28000 gamma: 0.1 weight_decay:0.0005 display: 100 max_iter: 28000 snapshot: 5000 snapshot_prefix: "snapshot/" solver_mode: GPU

my dataset is 10572 person，total 24308 pictures。its my own data，not the CASIA-WebFace data。

thx！

[haoranD]

in caffe-sphereface , i change the cmakefile.txt and makefileconfig, and successfully make all,but making test was failed,so can't finish the installation.but my own caffe is ok

help！please

src/caffe/test/test_neuron_layer.cpp:17:38: fatal error: caffe/layers/log_layer.hpp: can't find compilation terminated. Makefile:581: recipe for target '.build_release/src/caffe/test/test_neuron_layer.o' failed make: [.build_release/src/caffe/test/test_neuron_layer.o] Error 1 make:

[SeeU1119]

@haoranD I think caffe lack of log_layer.cpp &cu & hpp ,you can download caffe from BLVC,and add three files to src/layers and include/layers cpp and cu add to src/layers ,hpp add to /include/layers . I also get this problem and i take the above steps and succeed finally!

[HITxuefeng]

@haoranD I think caffe lack of log_layer.cpp &cu & hpp ,you can download caffe from BLVC,and add three files to src/layers and include/layers cpp and cu add to src/layers ,hpp add to /include/layers . I also get this problem and i take the above steps and succeed finally!

yes,it is useful.thanks