search

tornadomeet / mxnet-face

Using mxnet for face-related algorithm.
Apache License 2.0
544 stars 206 forks source link

face verification 10 times slower than in caffe with only cpu works? #12

Open BryantLJ opened 8 years ago

BryantLJ commented 8 years ago 
Hello, i tested your face verification code on my laptop. However, when i tested on LFW,  i found that it costs about **750 ms per pair matching** if only cpu was used, approximately 10 times of 67 ms, tested in this essay, '_A Lightened CNN for Deep Face Representation_', which used caffe to test in the same situation(only cpu worked).  
PS. my laptop's cpu is i7-3632QM, while the essay uses i7-4790, i think that there should be some slight difference, but the difference shouldn't be so huge.
So here is the question, whether it is because mxnet  itself is slower than caffe in this situation? or any other reasons?
tornadomeet commented 8 years ago

@BryantLJ hello, i do not test the cpu/gpu speed of lightened cnn model, when i have time i'll test it. this model is almost the same with caffe, another, there maybe some runtime diffience, but should not so big i think.

perp commented 8 years ago

@BryantLJ You must enbale OpenBLAS multi threaded. The time cost should be 1/3 of your current time cost.