News for hackers. Stuff that matters.

A comment on Tensorflow User Experience. Parent comment

Re:Tensorflow User Experience by m0zg on Saturday December 07, 2019 @12:44PM (#21731498)
Anything that can't use mobile GPU (or DSP/TPU for quantized inference) is pretty useless IMO, because it's just not energy efficient enough to be practical in a battery powered device, even if it's fast enough.
[ Parent | on HN ]
Re:Tensorflow User Experience by Nimitz14 on Sunday December 08, 2019 @03:21PM (#21738036)
Once pytorch is updated to use XNNPACK (being worked on right now) I think it should be fine to use. That plus QNNPACK makes inference quite low on power usage in my (admittedly limited, just integrated XNNPACK) experience.
[ Parent | on HN ]
Re:Tensorflow User Experience by m0zg on Sunday December 08, 2019 @04:15PM (#21738363)
As a rule, CPU burns at least 5x the energy per FLOP. So no, CPU is not a viable option on mobile if you need to do inference constantly. For "every now and then" cases, sure.
[ Parent | on HN ]
Re:Tensorflow User Experience by Nimitz14 on Monday December 09, 2019 @05:35AM (#21741604)
Interesting, thanks.
[ Parent | on HN ]