no code implementations • 18 Jan 2019 • Timothy Whithing, Thiam Khean Hah
We present a new methodology for improving the accuracy of small neural networks by applying the concept of a clos network to achieve maximum expression in a smaller network.
no code implementations • 11 Jan 2019 • Thiam Khean Hah, Yeong Tat Liew, Jason Ong
The projected performance on a multichip persistent implementation of all Resnet50 convolution layers is 10k im/s/chip at batch size 2.