Unrolling Ternary Neural Networks

Stephen Tridgell, Martin Kumm, Martin Hardieck +4

The computational complexity of neural networks for large scale or real-time applications necessitates hardware acceleration. Most approaches assume that the network architecture a…