face-everthing
face-everthing copied to clipboard
run in cpu?
first,thank you for your shared codes! now,I want to test the speed on the cpu,so i want to know if your code can run in cpu? @tpys
i have the same question:)
i don't test my code with cpu, but the detection alignment and recognition part, are all trained with caffe, so they definitely can run in cpu, but need to change some code. like caffe.set_mode_cpu(). current the most time consuming part is alignment. i am trying to retrain a small face alignment model to reduce that.
Do you test the code with cpu successfully? @billtiger @anguoyang