Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by modeless on Sunday November 10, 2019 @06:10AM
(#21497513)
Edge TPUs support 8 bit integer math only. Training is floating point. That's not a small change. https://coral.withgoogle.com/docs/edgetpu/models-intro/#mode...
[ Parent
| on HN ]
There are no comments on this comment yet.