Twitter | Pretraživanje | |
DeepMind
“Fast Sparse ConvNets”, a collaboration w/ [], implements fast Sparse Matrix-Matrix Multiplication to replace dense 1x1 convolutions in MobileNet architectures. The sparse networks are 66% the size and 1.5-2x faster than their dense equivalents.
Reply Retweet Označi sa "sviđa mi se" More
DeepMind 26. stu
Odgovor korisniku/ci @DeepMind
We also introduce a technique [] for training neural networks that are sparse throughout training from a random initialization - no luck required, all initialization “tickets” are winners.
Reply Retweet Označi sa "sviđa mi se"
Bharath 26. stu
Odgovor korisniku/ci @DeepMindAI @GoogleAI
Awesome Work DeepMind and Google. Are you going to open source the model?
Reply Retweet Označi sa "sviđa mi se"
Erich Elsen 26. stu
The models are open sourced: The kernels are also open source as part of XNNPACK:
Reply Retweet Označi sa "sviđa mi se"