Hironsan/anago

slower on GPU than CPU when training the NER model

eshijia opened this issue · 0 comments

eshijia commented

I find that the training speed is slower on GPU than CPU. I think maybe it is because the fit_generator function. Any ideas?