4 ms·
Do we know if this is a just a reference implementation or what Google uses in production? My guess the numbers will come down pretty quick. From what I can se
by blazespin 11y ago
Do we know if this is a just a reference implementation or what Google uses in production? My guess the numbers will come down pretty quick. From what I can see from mxnet versus theano versus torch is that GPUs are really the final determinant of speed and not the framework.